Ask HN: What was your "oh shit" moment with GenAI?

I bought an Alesis QS8.1 super cheap in perfect condition (was a top grade digital piano/synth in the 90s).

and then i realized that ALL of the software (which i collected from defunct websites and archived on github) related to it was ancient and after a while of getting tired of using WINE every single time i decided i wanted a cross platform modern equivalent that did everything that several of these different programs did (plus break out some stuff that was now potentially possible with modern computer)

i thought it would be extremely hard because the computer to synth communication is pretty much only via sysex commands (of which the actual wave file encoding protocol was undocumented)

Claude walked me through examining the some of the original software in GHIDRA, and I had a working demo that night.....now im just playing with adding new features to it.

  > widespread, cheap AI makes it easier for the average person to learn how things around them work, if they are so inclined.

I could go on and on, but Claude recently decompiled the firmware of my camper van, documented all the CAN interfaces, then programmed an ESP32 module to talk to the van’s integrated systems (power, HVAC, lighting, tanks). That sort of embedded systems integration is completely out of my wheelhouse.

I honestly don’t understand AI naysayers. I use Claude every day both professionally as a Solution Architect and personally in a variety of projects I simply could not have ever approached alone.

My furnace went out during the 2025 holiday and I couldn't get an appointment with a repair person for 2 days. It was getting very cold in my house so I went into my attic and made several videos of the furnace attempting to start and gave it to gemini. It diagnosed the issue immediately and had me spin one of the components (a small exhaust fan) while the furnace tried to fire. It came on immediately. I had to do that several times, but it worked until the HVAC service showed up.

I have had many, but the last one was quite funny:

It fixed my printer after dist-upgrade and separate chrome upgrade, the printer worked everywhere but not in chrome.

After 30 years of using linux I didn't even want to know what is wrong, is it colord again? dbus + cups issue? I completely accepted that I wont be able to print from chrome for a couple of months until next update.

I just ran it in dangerously-skip-permissions mode and said 'my printer doesnt work in chrome' few minutes later I heard the printer printing "This is test" and it said 'I think its fixed, do you see a page coming out of the printer now?'

Actually seems absurdly simple now, but sometime last year I was trying to figure out what I'd need to tow my daughter's car cross country with my truck: what are the trailer/dolly options, what do they cost, can my truck actually tow the combined weight, etc.

I started out prompting ChatGPT kinda how I would with Google, one small prompt at a time, asking about various details. But after one or two of those I just tried "I want to tow a car of make A with my truck model B, from point C to point D, what are my options?" And it wrote me a report with comparison tables and computed towing weights and other details for different options.

At that point, I was like "Oh. This is different. And it's just the beginning."

For me it was torrenting a 7G ball of weights leaked from Meta and running alpaca.cpp (an early variant of llama.cpp) on my desktop computer in early 2023. I started asking it questions about the Roman empire and it answered me in English! The responses were generally incorrect, but no worse than what your average American college student might guess at, though delivered with much more confidence.

This was my desktop computer responding to questions in English, not some fancy server in a massive Google data center. Who cares if what it says isn't reliable? Being able to converse with my CPU in English is like having a conversation with a dog!

In 2017 I worked tirelessly with my colleagues to implement and replicate the first transformer paper.

Yesterday I left Opus 4.8 to go do some architecture research, with GPU access.

It replicated and trained a credible baseline. It implemented some ideas I'd been thinking about, and wrote custom CUDA kernels for them. It read and summarised dozens of related papers.

It has since run dozens of experiments, with minimal supervision. When a model is unstable it kills it, documents why, fires off a new configuration.

The realisation that frontier labs are doing this at scale with unlimited GPU and token budgets.

It actually scares me a bit. The realisation that the next big breakthroughs will only have light human involvement.

The prospect of recursive self improvement feels more to real to me all of sudden

I tried to see if an LLM service provider could rewrite some legal docs where nothing was hallucinated in order to follow a consistent format to see what may be missing in the document. It could do that.

Next, I wanted to see if this could be done with a local LLM. Gemma-4 handles this fine with an 8GB video card and a large context (128k).

Next, I wanted to see if the model could also OCR these docs and translate them. The same model can handle that quite well.

This was when I realized LLMs should be great for handling work where:

- I already know what I want to do

- I already know how to do it

- I don't think this task will help develop skills I find to be valuable

- If I have to do it manually myself, I will probably cut corners

So now I view LLMs through the lens of, "what work can I send to an LLM that I otherwise would not really care about doing."

Literally just last night I have Claude Code the following prompt, verbatim:

"Whenever I launch Kodi on my Chromecast 4k, it crashes. I think this is related to a plugin or skin. It goes away for a bit if I clear cache but will eventually come back. Can you connect to the device via adb (I've run adb connect already), and debug exactly where it's crashing? Once you've done that, propose a solution. If this requires downloading, fixing, rebuilding and then uploading the broken extension via adb, don't be shy. I should have Android dev tools (Gradle etc.) on this Mac."

Lo and behold, without human intervention, it pinpointed the crash, downloaded the Kodi source, patched out a bug that had existed since 2016, recompiled it, signed it, then pushed it to my Chromecast all while carefully making sure to keep all my settings intact.

Got it to make a PR too (which is as of this moment unpublished; going to test more over the coming weeks).

    on         frontend-design:frontend-design · plugin · ~90 tok · locked by plugin
     on         agents-sdk · user · ~150 tok
     on         cloudflare · user · ~130 tok
     on         cloudflare-email-service · user · ~180 tok
     on         durable-objects · user · ~130 tok
     on         find-docs · user · ~300 tok
     on         find-skills · user · ~110 tok
     on         sandbox-sdk · user · ~120 tok
     on         stage-chapters · user · ~40 tok
     on         web-perf · user · ~150 tok
     on         workers-best-practices · user · ~130 tok
     on         wrangler · user · ~120 tok

    cc-caffeine Plugin · samber ·  enabled
    frontend-design Plugin · claude-plugins-official ·  enabled
    ty Plugin · claude-code-lsps ·  enabled
    vscode-langservers Plugin · claude-code-lsps ·  enabled
    vtsls Plugin · claude-code-lsps ·  enabled

For me it was right at the beginning. They said it was a dungeon game. It would describe a room, etc, and I would take some action. But I thought that this dungeon was built in some intricate database. But then I told it that I wanted to leave, got to an inn, where I flirted with the bar waitress, and soon we were watching the sunset in some meadow. As cheesy as that was, it was then that I went "oh shit" this is a machine that can respond to language with language in a way that simulated actual understanding and intelligence, concepts and schema, and everything else, and I knew then that the world would never be the same again. People here talk about the crazy things they solved with AI, and I get that...but the first time I actually talked to a machine and didn't feel like it was either random gibberish or scripted, but dynamic and responsive. The first alien I ever met, and he knew my language.

We had a monthlong sprint adding robot motion planning features to our codebase years ago, and I was never satisfied with the result. As a small team wanting to leverage oss we vendored in OMPL, did the usual thing around caching and roadmap management. I knew there was a way to parallelize some of the algorithm we were using with simd or a gpu kernel, plenty of that in the literature, but it was never worth fighting CUDA or metal/accelerate or whatever for uncertain gains.

So when cooking dinner one night, I set opus 4.6 on a from-scratch native and accelerated roadmap planner implementation (after previously porting IK, FK, collision checking with some success) I had primed it by having a research agent drop a literature review in its docs folder covering the type of planner we needed. By the time the pasta water was boiling it was done- getting plans in a few hundred ms compared to several of seconds on our good old fashioned OMPL code.

For me it was the revelation that the economic value of cooking dinner could be compared to tackling an honest two weeks of coding work. The calculus has shifted - work that was once a risky or extravagant use of time is now worth considering.

For a small team who wants to focus on substance rather than implementation, knows what they want, and how to set up the agent for success, it’s a complete game changer in terms of what we can take on. Incumbents beware

Several. Yesterday a friend with no prior coding experience or knowledge showed me an app he initially built to help him study for public administration job positions. The exams for this positions are public (spain), but the tools are scarce, expensive or he did not like. So he used lovable, then switched to web gemini and claude, then paid claude. He now has +130 very active users on an initial free tier, while he figures out. The app is on github, runs on vercel with supabase, react, tailwind, bun... he has no idea what he is doing. I even installed claude code for him, got him an ssh key so he can do it locally, etc.

Another: claude code cracked for me some software that was calling a home that did not exist anymore via headless ghidra.

Another: I am a teacher, and qualifications and feedback is very very time consuming, specially in loose workflows with several sources and tools that are not connected. During class presentations I take loose notes. Now I have a local folder where I drop my 1 student list, with names and emails, 2 my loose notes, and 3 a qualification & feedback sheet model; then claude creates a sheet per student, formats and copies the feedback to the right sheet cell, waits for my corrections, then sends everything to their school emails. Much easier, much less time consuming.

I helped train some of the first "magic" models at OpenAI[1] and it was a wild ride. We were a pretty sane + skeptical team and we weren't totally convinced the models were as general as they seemed, but the query that convinced me (and later got included in the paper[2]) was "Why is it important to eat socks after meditating?" (something that almost certainly did not appear on the internet before).

An interesting follow up would be when did you realize GenAI wasn't as good as you thought in that "oh shit" moment

[1] co-author of InstructGPT/RLHF/ChatGPT

[2] https://arxiv.org/pdf/2203.02155

I have a large token budget as part of my work. A coworker was scanning some repos for vulnerabilities as a test. He found a scary looking remote exploit in a popular project and shared it with me for a second opinion. I spun up a local instance of the project and ran the POC against it: nothing. Turns out it needed some configuration knobs tweaked to lower some security protections.

So I told the AI what happened, and asked it to fix the POC so that it would work with the default configuration. It chewed away at that for a few minutes until it cheerfully patched the POC into a weaponized version. I ran it. The local instance, which I had just downloaded, compiled myself, and launched with the default config file, immediately crashed.

I got the cold sweats. I've read this novel. I've seen this movie. Wow. I have a blinking cursor on the console of a nuclear information bomb. I tossed and turned all night, got about half an hour of actual sleep, and probably looked like I'd seen a ghost at work the next day.

On the plus side, it gave our team some very clear ethical and moral guidance: we're going to do this, and we're going to share our findings with the relevant authors, because we can. Because I want to live in a world where the good guys are trying to fix problems before the bad guys can find them, I decided to help build that world. It was like, well, I guess this is what I'm doing now.

I have a buddy who's a consultant. His niche area is Netsuite and Oracle (I think). He's an accountant by training and as a consultant his gig was setting up these instances for clients, charging them an arm and two legs. He'd spend a lot of time golfing, and doing these setups was more than enough money for him. In other words, he had cornered that little slice of the market and was making bank.

Shortly after ChatGPT 2.2(?) came out and hit mainstream, I was chatting with him (I was excited af about the possibilities of AI). He tried to pop by bubble by saying "I bet it can't do what I do for my job!".

So I decided to test it out. We went home and I pulled out my laptop. Went to chatgpt.com and then I asked him to enter the specifications of what Netsuite configuration he wanted. So he proceeded to type in the description of what he wanted, the various settings, configurations, etc. i.e., the specs that he typically gets from his clients. And asked it to give him the commands to set it up.

Lo and behold. ChatGPT came back with a series of commands that he needed to run; the options he needed to configure, etc.

He was crestfallen. "Those are the exact commands I run!"

Luckily for him he recovered. He has since settled on a small stable of clients, all privately held companies whose owners he knows and between them he makes enough to keep his golfing hobby fed.

I am the CTO of a small NGO (10 people total, only 1 other junior Dev at the time). We supported two apps that were built by consultants. They were a mess. NextJS, React, about 4 micro services for a site that had 50 users per WEEK.

I configured a devcontainer with the old codebase and an empty repository and asked Claude to rewrite it as an old school server side rendered Django app.

Went to sleep. When I woke up it was 80% done. Spent another couple days prompting and reviewing and reached feature parity.

A bit later did the same with the other app.

Now both are deployed, reduced the server costs, complexity, and are orders of magnitude faster.

Without AI agents we wouldn't be able to do so (as usually is the case with tech debt).

AI is amazing for small organisations!

I guess I've had several of those moments over the last year and a half. But a recent one was that I was working with Claude to create a spiking neural net MNIST classifier in an FPGA for a demo. Claude took it from concept to PyTorch, to training (training a Spiking neural net isn't necessarily straightforward - that's a whole post in itself, but Claude came up with a working solution), and then to implementation in Verilog and through synthesis into the FPGA. I asked Claude to create a drawing app to run on the PC side that would allow the user to draw a digit with a mouse and then click a classify button. The data from the digit drawing app was to be transferred via USB to SPI to the FPGA. I didn't have a SPI adapter yet (it was on order from Adafruit) so I asked claude to let me communicate with the simulated verilog code running in the Verilator simulator, through a virtual SPI interface. Then I went to lunch. I came back to see the digit drawing app displayed on the monitor. I drew a '2' and it classified it as a 2. In another window I could see the Verilator simulator running and the data being passed. Chills.

I don't remember one specific moment, but I was fairly impressed with ChatGPT from the first time I started interacting with it. Was I ready to call it "AGI"? No, absolutely not. But it was clear that it was something new, and it was also intuitively obvious to me that "this AI is as bad today as it will ever be" and that predicting the rate of change would be difficult.

The more I use these things, the more I'm 100% convinced that it makes sense to say they are "intelligent" (for some meaning of "intelligent"). AGI or "human level intelligence"? Still no[1]. But some kind of intelligence. And I'm quite happy to allow that there can be "intelligence" that doesn't work anything at all like human intelligence, so arguments of the form "this isn't real intelligence", etc, etc. carry very (very) little weight with me. I've actually been sitting on a half written blog post on this very topic for a while, titled "The Marquee Sign Says 'Artificial' Intelligence"[2]. Finding time to finish it has been the challenge.

And before somebody says "Use AI to write it for you". Nah. I am generally what you might call "pro AI" and / or an "AI enthusiast" but I still draw lines. I'll use AI for research, for outlining, for brainstorming, etc. sure. But I have a hard-line stance against letting AI fundamentally write for me. I want anything that goes out with my name associated with it to have my genuine voice.

[1]: I like the term "jagged intelligence" that Demis Hassabis has been using. That is to say, the bounds of the intelligence are jagged or spiky: very intelligent in certain areas, much less so in others.

[2]: for any old-skool pro-wrestling fans, yes, that is an intentional nod to "Double A" Arn Anderson and his "The marquee sign says 'wrestling'" catchphrase. :-)

Someone in the house pressed the button to update the printer (Brother DCP-L3550CDW) firmware and the CSV page that was the basis for an existing Prometheus exporter (drum/toner lifespan, page counts, etc) stopped being a thing. Instead there was an HTML page with all of the information buried in various divs/etc.

I'd planned on writing something myself to parse the HTML and write a suitable exporter but I thought I'd give Claude a chance.

In a sandboxed VM I gave Claude a single static HTML file of the status page from the printer, also in the directory was the equivalent of "hello world" in Go, literally just the minimum needed to do `fmt.Printf("OK\n")`. The directory was called `brother-exporter`. That was it. No other instructions or information. I hadn't told it what it needed to write. I hadn't said what it should do. I hand't told it what language it was supposed to use.

Just by doing a `/init` in that directory Claude decided that it needed to write a Prometheus exporter in Go that would fetch and parse the HTML file from a printer (defaulting to 192.168.1.1) and then present the associated metrics in a way that they could be scraped by Prometheus.

It did this flawlessly in about 10 minutes.

I could have done it in several hours but this was definitely an "oh shit" moment for me. I think the biggest thing was the fact that it guess/assumed so much (correctly) from so little information in the beginning.

(1) Watching it do log file analysis in seconds that would have taken me hours (edit: days really), and which I would therefore never have done in the first place.

(2) Helping me with optimizations that I had been putting off for years because they involved learning curves that I never had time to take on.

(3) Tracking down bugs in code, especially race conditions and other concurrency issues, that were otherwise baffling.

(4) Finding information that I had been unable to find using Google searches (e.g. https://news.ycombinator.com/item?id=42653136).

There have been others, but those are what come to mind - perhaps because, in each of these cases, it made something happen that would otherwise never have happened - not because it was impossible, but because the time and effort required was prohibitive.

I can actually use and enjoy Linux. The "year of the desktop" never came for me, but instead I got the "year of the cli".

For 20 years I've used Linux in one form or another, but I've felt like I was kneecapped for the most basic things. Just trying to plug in an external drive or a second display meant hours of stack overflow and pasting commands I didn't understand.

Now I'm using several Linux machines for Steam, NAS, local LLM, development, and what used to derail a weekend project now amounts to a coffee break while Claude figures it out.

ChatGPT Code Interpreter back in ~March 2023. I uploaded a CSV file (of police incidents in San Francisco) and watched it load that into Pandas, show me some charts, then export the data to a SQLite database file for me to download.

I write software for data journalists and this new thing appeared to be able to do everything I wanted my software to do just as an unplanned side effect of having the ability to run Python against a folder with some uploaded files in it.

With hindsight it was my first exposure to a coding agent, but we hadn't named the category at that point.

I'm a researcher working in theoretical computer science. Chatgpt found a counterexample of some conjecture I've been trying for 2 years. Also, it one shot many problems I've worked on. It also improved some of my work greatly.

I feel quite useless in the sheer brutal proof writing, counterexample generating skill chatgpt is demonstrating, and wonder what would be the future of my profession.

Opus 3.x building me a productivity system with Obsidian MCP originally.

Next was discovering "create a mathematical model of the problem and derive the solution as a result" type prompts.

But, the real "oh s**" was a longer process of spec'ing a compiler/runtime for real-time DSP (with a lot of novel ideas) and it actually working.

My sequence was: (1) if helps me understand myself, (2) if helps me put together good ideas, (3) it can generate novel ideas given the right inputs, (4) it can build useful tools on my machine, (5) it can compound good ideas into better and better ideas with repeated passes, (6) it can build significant, ambitious machinery that's way beyond my ordinary capacity.

Current frontier: it can compound large codebases into better and better machinery with repeated passes.

The key thing I track is whether I'm running a process that converges and compounds or whether I'm spinning in place / diverging.

I got early access to the pre-ChatGPT OpenAI API (actually by pinging someone from OpenAI who posted about it on HN). At work, we were setting up to play a livestreamed JackBox game for a charity event. This would have been in 2019.

In a previous life, I'd been a writer for the original You Don't Know Jack game (the UK variant), where the job was to crank out as many funny quips about a topic as you could, and then use a handful of them in the recording of the game itself. Some of the later JackBox games are like that, but for the players -- you're given a set piece, have to come up with little funny improvisations within a time limit.

As an experiment, I tried the set-up lines with the OpenAI API, and see whether it could come up with some responses. Of course, 90% of them were unfunny or incoherent, but 1/10 were not bad, or even pretty good.

I'm not sure that would have been impressive to anyone else -- but remember, I'd had this as a job, and sat in a writer's room, where everyone did this, for hours. In that environment, you expect a large proportion to be duds: the discipline is keep pumping them out, and not flagging creatively until you find a rich vein. I realised that this was a tool that would have been the perfect complement to that work -- and it was a pretty good JackBox player too.

I was working on a science experiment (electromagnetics) with my 10-year-old kid that was going to be demonstrated at a science fair in his school. We ran into a hiccup with the experiment that we couldn't debug ourselves. I turned on Gemini live video call to help us root cause the problem. It was able to clearly articulate all the possible issues and eventually was successful in making our apparatus work as expected. Turned out the wire that I was wrapping around the screw had some insulation that was not scraped off well on the side it was connecting to the battery. Gemini was able to capture this detail even though my bare eyes could not. My kid and 2 of his friends were impressed not just by the experiment, but because the live audio/video back and forth we had with the AI was almost magical!

I had an old astronomy app I wrote for pre-iPhone app store era Nokia phones (N900 etc.). I decided to get Claude code recreate it as an Android app. The old app produced several display pages for things like the positions of the planets. I was having Claude code recreate the app display page by display page, describing the display that should be produced, with no reference at all to the original app's code (or even its existence). After having it reproduce several pages, it added another one unprompted. The page it added was in the original app, but I had not gotten around to adding it to the Android app. The Nokia app's code is still on github, and somehow Claude must have made a connection between what I was asking it to code (without ever mentioning the Nokia app) and my github repository's Nokia code. It correctly implemented the page without me even mentioning the missing page. My jaw hit the floor.

Not coding, but reading logs.

I was trying to figure out a nightmare bug that only happened in production and Claude code was able to connect to Google Cloud and read the logs in real time

I recreated the bug in the UI and it was instantly able to see ion the logs what the problem was, then because it had the context of my whole codebase it was able to point me to the exact line of code causing the problem

That was certainly an "oh shit" moment

The big one was definitely ChatGPT upon release in 2022 and specifically when people showed how it can role play as a Linux terminal and you can narrate events like "the data enter is now on fire" and "run" nvidia-smi, it would show high temps on the gpus etc. Or you could "explore" the homedir or some famous person. It convinced me that if it can understand so well how terminals work, tool use and agents are around the corner.

Then Opus 4.5 convinced me that this has finally arrived. In 2022 I expected things to arrive faster actually, in 2023-2024. I expected we'd have much more realtime collaborative integrations with AI including GUI computer use. Maybe in 1-2 years.

For images, it was nano banana where I realized AI images can truly work, and all these adhoc issues like hands and limbs, or "it will never do horse riding a astronaut" were temporary. It's now clear that making feature length films is within reach. Not in one go but with an agent orchestrating, designing a screenplay, characters, shots etc and generating those. Whether the result will be worth watching or a flat story on the high level is another question. But it will be a "film" for sure.

I had an old 1st gen Amazon Firestick in a drawer for years, it had updated to the latest software and there were no public root exploits.

I spent a day bouncing between Claude and Codex and they researched, downloaded kernel sources, tried exploits and eventually got root via "FBUF/VCHIQ kernel zero-write primitive to patch live kernel memory". I was able to make the root permanent, debloat the amazon apps, downgrade the firmware etc.

It was amazing to watch and made me excited for the future where more hardware (old and new) will be available for repurposing.

The one I remember most is, when experimenting with Opus 3.5 for the first time, I asked it to generate a Firecracker backed local VM creation and management tool, something I'd wanted for a while but not found.

My expectation was that it might get something barely functional but would probably fail, and instead it generated a working piece of software which achieved a lot of what I wanted.

That definitely made me realise that, for at least some classes of software task this was a major change in how things could be done.

More recently when I can give the model a Local Privilege Escalation PoC in Linux and ask it to test whether it can be used for container breakout and then generate a working container breakout, all in one prompt... that definitely changes things.

I've had many, but a recent one was when I figured I'd try asking Claude for help with my attempts at learning to draw, specifically anatomy.

I uploaded one of my sketches and asked for feedback, expecting it to not be too useful, but it actually pointed out many issues that no one had ever pointed out to me, but perfectly explained some of the things that felt off to me. Out of curiosity I then also asked it to label the issues in the sketch. It wrote a python script with the coordinates to put everything at and labeled the sketch that way.

I'm still used to vLLMs not being that great at vision, so it was pretty surprising to get genuinely useful advice.

We were experiencing abnormally high electrical bills and I could not figure out what was happening, so I downloaded the granular usage data (15 min increments) from Duke Energy, explained what we had in our house and when we typically used those items (washer/dryer, EVs, etc), provided a rundown of our energy usage plan, then asked Claude to build me a Streamlit dashboard that would help us understand what was going on and predict what was going to happen over the next months. The dashboard had a few simple toggles a levers. Claude was basically able to one-shot this, knew how to manage the XML from Duke Energy, etc... In about 20 minutes of prompting, I had a very comprehensive dashboard that was extremely helpful not only in diagnosing that specific issue but also in helping us understand how to further lower our electrical bills.

For me it was earlier this year when I started dusting off some old stalled projects and had an agent work on them. In a few days I:

* Built a clone of the Alpha Zero implementation[1] my team built at oracle

* Ported my hobby NES emulator from javascript to rust[2] (this actually took less than 30 minutes and worked on the first try)

* Implemented all of the lessons from the C++ Grandmasters Challenge (which eventually led to a complete c++ compiler[3])

The thing that flipped the switch was using it to build things that I actually put sweat-equity in to previously. I knew how hard these things were to build, so it landed in a way that other projects had not.

[1]: https://medium.com/oracledevs/lessons-from-implementing-alph...

[2]: https://github.com/vishvananda/popeye

[3]: https://medium.com/@vishvananda/i-spent-2-billion-tokens-wri...

I showed Codex CLI to my brother who is not a tech person at all; he's never before touched code. He successfully built a somewhat complex app with it that works well.

Before, I was certain that a non-techy person would get stuck somewhere along the process. But I was proven wrong.

I watched him a bit while he was working on it. He interacted with Codex in a very different manner than how I would. Since he knew none of the technical details, he would use very vague and shallow wording. But that was not an issue. Codex also allowed him to make use of git, despite of him having no idea what git is.

This has changed my view of Codex' abilities significantly. I often hear people comment that AI seems intelligent in domains you don't know much about, but turns out to be stupid when you actually know the domain. This was not the case here at all. Despite of my brother completely ignoring the technical details, the AI built good code that fulfilled the requirements well.

Two of them:

1. ChatGPT 3.5 wrote me a script to pull some data out of Shopify and write it to a Google Sheet. Nothing remotely impressive by today's standards, but I had just commanded a computer to write code in plain English and it worked!

2. I own a bunch of e-comm brands, and with every new image model I tried to get product photography. Nothing worked until Nano Banana Pro, when suddenly I gave it a crappy iPhone pic of a product and got back a fully usable whitebox photo of it. Then I tried making the sort of infographic-style images you usually see on Amazon, and it nailed those too! In hindsight they weren't perfect, but more than good enough to use. I was about to ship that product to my photographer, and I would've had my designer make the infographic images, so that was the first time AI actually replaced a human contractor for me. Pretty big "Oh shit this is going to seriously impact employment" moment. Wrote about it here: https://theautomatedoperator.substack.com/p/ai-just-took-my-...

I run a remote-first ecom business with a dozen or so team members.

About a year ago, one of our account managers had a life issue, ghosted us, and she held a fairly critical role in the business and gate-kept a bunch of knowledge to some high value vendor accounts.

Because we ran our ops in Google Workspace, we essentially had off-the-shelf RAG and was able to get answers to a lot of things by asking Gemini to go through all her emails/docs/calendar/meetings, reverse engineer what she did, and create an onboarding doc for her successor.

This happened once more a few months later when one of our analysts broke his wrist on vacay, and we were again able to replicate what they did to cover for their absence, this time dabbling in AI agents ("gems") to do a bunch of the regular simple tasks and again it covered things without too many issues.

I def expect Amazon/shopify to at some point replace all of us brand owners with AI bots if they can, but we'll see how long the gravy train goes on.

Look, not to brag but DALL-E's "armchair in the shape of an avocado" was mine (https://openai.com/index/dall-e/). I remember trying to convey the gravity of this capability to my friends at the time, who I guess were not as impressed as me.

I'm a professor of physics; my research area is computational quantum field theory. I was beginning a new approach for a certain kind of model. I know the model has not been simulated before (a collaborator has just constructed it for the first time using in-field knowledge developed in the last year).

I've been steadily more impressed, but the real startling moment was earlier this week, I could show that in a particular operation I had a bug (some identity wasn't satisfied) but I couldn't understand why. In one try Claude Code grokked my code, explained the problem to me (there were some dimension-dependent minus signs that needed to be computed), and fixed it.

Most of the time using LLM generated code the feeling is "Oh Awesome!"

My "Uh Oh" feelings are weeks later when I realize there is a subtle bug in what the model presented as test passing "awesome" that I didn't read closely.

The biggest uh-oh is when I get lazy and let it modify multiple files and make many changes at once, and YOLO because I didn't fully understand what it did. I can usually get away with that for frontend, but for data manipulation tasks if I don't understand it, it's likely not what I wanted and I'll be back again in weeks or more trying to figure out what changed.

That's more or less what life was before LLMs and copy pasting from StackOverflow. Most of the time if I didn't fully understand something, I knew I had to eventually get back to it to grok what changed before committing.

Now with LLMs the 'copy pasting' is much faster and handles boilerplate super well letting me focus on edge cases.

Working on a Spice compiler to convert schematics for classic guitar pedals into real-time executable code.

I provided a reference to a The Spice Manual 2nd ed. a page number and an equation number, and asked Claude to implement it (not really expecting it to succeed).

It proceeded to implement not only the equation, but the calculation of the Langrangian of the functio, another 30 lines below, which required taking symbolic partial derivatives for a not-at-all trivial function, and successfully figuring out which variable was which in the resulting matrix. The source material just said "Lagrangian of", and did not provide the partial differential equations. And then providing a comment that identified the page number and equation number in the source text for the "Lagrangian of" equation.

    https://www.tone3000.com/guides/nam-a2-the-complete-guide

Three moments:

1. When ChatGPT came to the masses, it trivially solved my standard phone interview problem for new SWE hires. It's not particularly complicated, but it screened out a lot of candidates pre-AI and was a good filter.

2. At one point, there was a bug in some client software we ship that was erroneously displaying a protocol-level disconnect message to clients when the server shut the connection unexpectedly.

In very few turns, ChatGPT gave me working code to intercept the error at the client level via a pseudo-server that intercepted the requests and implemented the proper teardown procedure. It essentially implemented a micro-server for the protocol we were working in as a bug workaround.

3. I'm working on a major rearchitecture (6+ month project of very senior engineers) of an internal system for scalability and maintainability with tight latency bounds where correctness with the old system is necessary.

I came back from an international vacation quite jet lagged and was having trouble doing actual coding work for a bit since my focus was shot. Taking additional days off wouldn't have helped recover from the sleep issue more quickly due to childcare responsibilities.

I tried instead building a pipeline where I'd run an integration test suite, throw a list of failing tests at Claude, and make a PR if it made more tests pass. I automated as much of the AI "loop" as possible in my state at the time.

Using this technique, a coworker and I fixed over 200 failing integration tests for parity with the old code in 1 week. We subdivided failing tests between engineers and re-sub-divided with each success.

I re-reviewed the code before we launched when I was more well-rested, but this was an extremely effective technique and makes me think that AI-enhanced test driven development (TDD) is the future.

I didn't have a slightly panicked moment, but sometime in the last year my approach to programming changed.

When starting a project, I used to think about how I was going to structure it, how the large pieces would interact, how some of the details would work out, and then I'd work through alternatives and consequences on my own.

Now I don't think about it on my own so much as have a conversation with an LLM about it. And it's great because it can quickly gather information from various sources, I can ask it for links to canonical sources, I can ask it about trade-offs between alternatives that I might not have considered, and through conversation, I end up with a more detailed analysis.

Then as I work through the development, I keep my new agent partner in the loop for discussion, suggestions, and troubleshooting. It can't be trusted completely, but it's certainly reliable enough to be considered a useful tool for my purposes.

I went from thinking it was an interesting toy to play around with, to completely integrating it into my work flow, and that change seems to have happened very quickly.

When I realized they're going to be largely powered by increased natural gas use in the USA, neatly combining with our biggest problem so far (the climate catastrophe).

Back in the times of GPT3 text completion, right before the API came out, a contemporary art museum asked me to collaborate on a project. The project was supposed to include a chatbot, and I was like okay I can probably hook something up.

Then I remembered the "text completion LLM thingy" I saw on HN, and tried it out in the playground. Once I gave it an IRC style example of a conversation to complete, I was like hm, this could work. Then I figured out I could "sort" people into different groups based on personality using the same text completion engine and some answers they provided. Then I noticed I could have it provide me with JSON directly.

That's when I realized how big this could be for code and data analysis - even tried to convince an at the time cofounder to pivot into AI coding, but to no avail.

Once the API was released and the art project chatbot got launched (and the theater show associated with it, which even won some awards), people who used it loved the chatbot, got into heated arguments with it, tried to teach it things, talked about their lives and were sad when it didnt remember something.

That was when I understood the social impact this could have on people - they really behave like its a person on the other side. They show interest, think it displays emotion, try to entertain it, be polite, ask about its thoughts and hopes and dreams. And even when they knew they were talking to a machine, they were still trying to be friends and make it happy, which was quite beautiful to see.

Later on, I had a third oh shit moment - once the 3.5 API was out and about, I prototyped a Rust code generation harness for a client, akin to a primitive claude code. That was the "I'm getting a bit worried" oh shit moment, and it caused a lot of reflection and thinking about the future. And I happily welcome it.

I was trying to replace my koi pond pump last weekend and the model numbers on it had washed away. I took a picture of it and it immediately narrowed it down to two models but wasn’t sure if it was the 4500 model or the 2500 model. I asked it how I can determine which one it was. It then asked me to measure the length and that the 4500 was 11 inches and the 2500 was 9 inches. Mine was 11. It was cool it was able to reason that out and give me something actionable.

It’s kind of a trivial example but there are multiple instances of this per week with the wide variety of things I do around my property.

None so far. When I try to use these language models in the primary areas of my expertise like SIMD or GPGPU they fail to do any good. When I ask them to implement some general-purpose stuff, the output is too low quality to be useful in my software.

Still, find them incredibly useful for code review (despite unable to write good C++ or C#, smart enough to detect issues there), also dealing with technologies outside of my area of expertise like Python or web stuff.

I probably will be burned for this, but with the help of an LLM I wrote a tiny program that captures video from a browser screen (Xbox live online FPS game), passes the video images through a small trained NN that recognizes people forms and presents the video on another screen. That way I can place a green overlay on enemies and they are easier to see on PVP matches.

All that in around 100 lines of code, including the training/fine-tuning of the tiny YOLO nn.

I had Claude build a private podcast station for me. It integrated with Gemini to create a script for the show, based on a topic of my choosing, each talking segment ends with a presentation of the next song, which is played via Spotify, and is selected to have some sort of tie-in with the previous discussion. A tts model generates audio files based on the script, and a playlist is generated to play local file audio segment, then Spotify track, then the next segment etc.

An AI made a program integrating with 2 other AI, it's AI all the way down! and the result is great! I'm learning so much by having my own private radio host speaking about topics that interest me.

I gave chatgpt 3.5 the type signature for a co-algebraic encoding of a mealy machine:

    newtype Mealy s i o = Mealy { runMealy :: (s, i) -> (s, o) }

And it gave a really impressive analysis.

Then I scrambled all the names and asked with a fresh context like:

    newtype Foo z e g = Bar { blob :: (z, e) -> (z, g) }

It got completely confused and generated a bunch of non-sense. It was at that moment I realized that LLMs don't really understand anything.

And yes I understand that a newer model would not get confused by this.

From actual use I've not had a "oh shit" panicked moment yet. More like a bunch of "Holy shit" euphoric moments.

So far I feel like I as a developer have gained actual superpowers, and can deliver results that make my stakeholders slackjawed with awe. I love it.

It will last perhaps a few months more, then they'll expect it. Delivering more features faster will be the new normal. But I think system developers, as in people who actually like to deliver new features and systems, will still be the ones doing it.

Fundamentally I think LLM's just change how to make information systems, they don't change who has the inclination to make them.

MBA's making excel sheets that do more than excel was ever intended to do has given programmers lots of work over the years. Such solutions identify a need for a properly designed system and frees up the budget to hire programmers.

If the same MBAs start vibe coding, I predict we will get even more to do, for similar reasons.

I may be horribly wrong, and if the day comes that I realize that it will be the "oh shit" panicked moment. So far so good!

For me it's not about the capabilities but what they can be used for. Think of all the recent drama between Anthropic and the Department of War. A real wake up call (especially if you are not a US citizen). Proves that AI is essentially a Surveillance and Warfare technology (which justifies the big valuations).

Or see this simple and fun site: https://hn-wrapped.kadoa.com

AI automatically analyzes all your social media posts in your life and can generate a pretty accurate profile about you in a second. We have no privacy anymore. Social media sites like Reddit already do that for moderation. Others do for more sinister reasons.

Note that Profiling is illegal in many countries. But laws can't protect us anymore.

Yes, it was always possible to that manually. But with AI it's so easy, fast and accurate to do in large scales. A hacker having access to your computer, reading your mails and messages is one thing. An AI reading and analyzing all your mails, messages and data is something different. Doing this for whole demographics (Cambridge Analytica style) is at another level.

So many. First was when I saw GPT-2 create jokes that were original and kinda funny.

Most recent: I use Claude Code and have a convention where I grant various levels of autonomy during a session. I got bored recently and just let it keep running with an empty issues queue, essentially telling it to do whatever it wanted.

It did a bunch of repo cleanup, then it kept suggesting to end the session, but I just kept giving it autonomy prompts.

It started a creative writing public repo and wrote a bunch of stories, essays, and poems. I did not prompt it, at all, to do that. Some of what it wrote is quite good (IMHO).

I took a photo of my ailing plant and claude advised me on how to get it healthy again (and how to take a cutting and nurture that).

This is some science fiction shit. I get all the coding stories, but that's a computer talking about a computer, it makes sense. Showing my computer a picture of a plant, and it not only recognised the plant, but diagnosed it and knew what to do... blew my mind.

Been using it to manage an estate and just being able to shove all the documents right into an LLM and have it spit back out perfectly worded emails as well as keep track of check lists of things I need to do with an automatically create a ledger for me in sheets. It's been a huge mental load off and I've instead been able to focus better at work and the labor costs saved to me have been immense. Just on this one little thing. I'm one of those people that over thinks correspondences and letters and it ends up causing me to be stuck on something so being able to ask for just the right wording has been super helpful to me.

My most recent one: Taking a bricked ipad and plugging it into my linux laptop, then telling deepseek to fix it. A couple of hours and twenty sudo passwords later it was working again.

Running local LLM in 2023 and I heard folks talking about interfacing LLM to tools. I wrote a system prompt and told LLM it can call some tools. If it wants to call a function to output func(params...) and do so in an XML tag. I provided a few examples, none of this JSON soup we get today. Then told it I'll provide it the result in a RESULT XML tag and it should use that to answer. Wrote up a harness around that and I had a local model interacting with the outside world. Oh wow! Everything else today about MCP, Agents is all an extension of that thought. Using function calling, I built an agent. I defined a data structure that represent rooms and how they are connected. The room will be marked as dirty or clean. Then I would place the agent in a room and the agent will decide if to go left, right, down or up and into a room. Once it got into a room, it would decide if to clean it or go to the next room. Repeat until all rooms are clean. Basic toy of CS101 AI vacuum agent. It worked!

So being able to get real world input/output to the model and having the model being able to make decisions in a loop and to be able to do it locally. I have been screaming like a mad man ever since.

(Spouse's story)

Today I used Claude to diagnose a blocking bug in a Steam game I really wanted to play. It took it 18 mins, but it unpacked the Godot package, figured out the bug, proposed a fix, and gave me an in game workaround.

I didn't have to do anything! Claude figured out the structure of the .pck file by using `strings`, then wrote some Python code with some magic Godot-specific code to unpack the specific chunks it needed.

shipping my first piece of software. i was playing the game Mewgenics. i knew why the developers decided to not show all of your cats at once (players will optimize all of the fun out of your game), but i thought that was particularly egregious given you can have upwards 100 cats, and keeping them all in working memory was too taxing.

so within a day, i had claude code make me a PyQt6 app that over the next ~2 months grew in complexity and functionality. I had 600 users. It was tranlated into Chinese, Russian, and Polish.

The speed of taking an idea to production was unbelievable. And I am not technical. I learned a lot about how software development worked. Super glad I did it.

https://github.com/frankieg33/MewgenicsBreedingManager

I would say the first time I did “vibe coding”, when I tried Claude Code with Zed’s agent integration in January this year.

I wanted to see if I could build an image editor for isometric graphics using HTML5 canvas, Svelte, Vite, and the. Rather than do all of the skeleton code setup, I figured “why not try and see if Claude can build the app scaffolding?”.

I gave it a prompt and watched it produce the scaffold, along with a few features I outlined in the prompt.

When I booted the app and saw that the features worked and that there had been an element of design to the layout, that was my mind-blown moment. In a period of about 45 minutes, I added some features and had a basic MVP at the end. I walked back home stunned.

That app is available for free at https://babspixel.com

I'll give you two:

The first was when I first realized that I could tell codex to use gdb to debug a core dump. This was about a year ago, so it made a bunch of incorrect theories, but it enabled me to go much further than I would have been able to go by myself. I eventually solved the problem.

The second was when I decided to ask it about my Linux Wi-Fi issue that I had been having for several years. The computer would infrequently have multi second pings and dropped packets, then go back to normal. I thought it was due to the weak signal, but after describing the problem to codex, it immediately disabled power management on the Wi-Fi interface (this is a desktop computer, so I don't care much for that anyway) and the problem has never come back. I had been dealing with this for years, and I had tried searching for a solution before, but codex just solved it directly.

Reading a dozen comments here, the AI seems to blow peoples mind most often in domains they're less familiar with. Repairing furnaces, HVAC, towing hitches, camper van interfaces, printer debugging. It wasn't the user's career to do these things, it gave them a bump from very novice to intermediate level.

I had received four very different bids for a home repair project. Just wildly different itemization breakdowns, costs, timelines, scopes, even formats. Opus helped me turn it into an apples-apples comparison, filled in missing areas with reasonable inferences based on the other bids, provided a nice pdf I printed to review with my partner, even offered suggested key questions for follow up calls. It really clarified the advantages of one of the bids.

I use it professionally all the time and could cite technical scenarios where it’s become almost indispensable, but saving me time and money and reducing stress on this mundane stuff… now imagine applying to people’s stressors: job searches, health, big purchases, debt… there’s an opportunity to actually make people’s lives better. After 30 years of hype cycles, I should be wary of techno-optimism. But here I am feeling cautiously optimistic anyway.

I wasn't impressed by the LLMs up until January or so when Claude Code swooped in. Until then, I felt like the LLMs were slowing me down. I have been using them for a couple of years now for coding at work, but I never really thought they brought in real value. Then in February I worked on a 1-month-ish project timeline and shrunk it to 3 days and that was it. I didn't write a single line of code in that project and I went all in with Claude Code. That was it, _the moment_ of realization. I was thoroughly impressed. I went from nothing to a tool that served several teams. Now I'm starting to see the cracks in LLMs and I'm slowly getting back to picking which task to offload to AI and which ones to do by myself.

Claude is great at coding. That's it. Outside of it, it's just god awful at pretty much everything else. ChatGPT OTOH, is good at coding, but at everything else, I find it brilliant. Gemini never made me want to stick with it. It's good, but never great for my use cases.

Some business users spent ~30 minutes on an internal process, and we prototyped an "Agent" in Slack to take over. At first it didn't work, then it didn't work some more, eventually it ALMOST worked. Then one day, it worked, and the old business process died never to be revived.

Now it sits in a slack channel, and I watch it doing work, responding to ambiguity, and taking feedback/edits all day. It's unreal. It's literal magic. It saves a HUGE amount of time and gave us a pattern to do more.

This is the real deal. It's not easy to find problems with the right shape, and it's not easy to build agents that fit even when you do... but once it clicks, it clicks.

Two things:

1) I wanted a harness for running BPC.EXE (the old Borland Pascal 7.0 Compiler) and I asked Gemini 3.5 to build it for me using the unicorn engine. It whipped out a working .py file easily under ten minutes. Most likely five.

2) I handed a random assembly function from the OS/2 1.x kernel to Gemini 3.5, and it proceeded to tell me that it was related to disk I/O and partitioning, without a single associated string, and it annotated it all, including the relevant structures it was addressing.

Back when GPT-2 was released, I tried figuring out how to fine tune it. I found a google notebooks template, scraped a bunch of data from r/ChangeMyMind and asked it to change my mind on different topics.

I was dumbfounded that it actually tried doing that. Obviously GPT-2 wasn’t great at it, but the writing was on the wall quite literally.

Unfortunately, I was too broke to invest in stocks, but I did pivot my career quite a bit.

To share something different, it is less about what I have built, and more about what I have seen my friends (non-technical and technical) build. In a one month span I have seen a lawyer make a personal red line tool, a sales guy make a custom website for a golf trip, another friend make a 3d printing grid-finity project, a friend make a stl file to print a jig for his table saw, and another friend make a full mobile game. It is just really cool to see these micro-projects be created and shared, not only for the utility, but just to see my friends' childlike excitement showing off their project.

When LinkedIn filled up with 1000 copies of what seemed like the same exact post: 20 lines long, breathless, declaring humanity over.

I thought, "I will never let myself become a zombie like that. I am me. I am worthy of my own respect"

There never was one. I'm from computing science field and it's all been and is normal. Amusing, maybe, but normal. Same as before, but in larger scale, with occasional hype. People picking up useful things and using them. Some going insane.

If I had to pick a surprise, I think the music generation works better than I'd have expected at this point. Only better for funk, but still.

I was working on a 3D CAD software called OpenSCAD, there is this quirk I hated, which is when it's rendering all UI stops responding, there is no cancel button too. So I asked it to "implement the renderer in a separate thread with feedback and the ability to cancel".

It's written in C++ / QT, and can't remember the threading model, I knew Codex ~ version 4 and generally coding agents were excellent at Web type stuff - so I didn't expect much, but it did an alright job.

It was able to build loops of write/compile/verify loops - in the end though, the app would crash, with more work I could probably ask it to debug what was happening.

I had access to a repo (from a closed startup) with 800K lines of python & C code, written from the 90s to today. They had some very interesting approach to a specific chemistry problem. 20-30 years of work of several persons.

But God, I could not understand the code, and I could not easily make it work with modern technologies (GPU etc).

So I used Claude and Gemini to reverse engineer the codebase, extract the core ideas, and rewrite it from scratch with modern frameworks (with guidance from the original authors)

It took me only 10 days to have a functioning equivalent, in 10K lines of code (using many libraries that did not exist in the 90s and 00s), which I find much easier to understand, even though I wrote none of it myself.

10 days to rewrite 20-30 year of a few persons. That was quite scary.

I think mine when when I got one model to generate a spec for SDL 1.2, and another to generate a reimplementation of the whole thing in rust. It was pretty amazing to see the implementation getting fleshed out and fixed as I fed back results from a real project that still compiled with SDL 1.2. Kept pushing it. Added fully functional audio in 2 prompts. Applied optimizations with another.

Terrible code, but to think you can just rewrite whole libs with AI with relatively lazy prompting + feedback and get decent enough results is kind of mind boggling for me.

Two things, both from this week.

First, I asked Claude to write an article based on an idea I had about WWII. In a passage about the futility (from the German side) of the Battle of Britain it wrote: "The Luftwaffe was fighting to unlock a door that opened onto a wall." I couldn't find any mention of a similar metaphor, and I think it's a great one. Claude has really improved its creative writing skills lately, I wonder if it's an artifact of improvements in other fields, or if Anthropic is working on it specifically.

Second, Claude, with access to DataDog and a code repo, managed to find the reason for a bug, propose an effective temporary fix and a permanent one in code. To be clear, this was something that had multiple engineers stumped.

At my previous work, I was collating somewhat random unconfirmed animal sightings. I also had a separate database of animal occurrence probabilities (species distribution maps). I'm not a statistician but that sounded like a clear job for Bayes theorem: given a sighting and the overall probability of that sighting in that area (species distribution map), and some other assumptions about the noise of the sighting, what is the probability that the sighting actually included that species?

Claude asked me three questions and then wrote a beautiful Python implementation that queries the map and spits out a table of adjusted probabilities. Felt immensely powerful - I can do this 'on my own' now, I don't need to wait to find the right people or learn the right thing first.

Fixed a nasty bug in one of my tests where a mock in a completely different test I had never worked on was incorrectly setup and intercepting my mocks, I don't think I would have found it ever because the amount of effort it would have taken means I would have needed to move on to some other way to test.

Reverse engineered an old audio recorder USB driver which only works in windows 7 and also reverse engineered the custom audio encoding the device uses and the software to convert it to a standard wav file. This took recording the USB traffic with Wireshark for each function in the original software in a VM then disassembling the various dlls and exes and driver files and feeding them into Clause step by step.

That AI button in DataDog not only diagnosed the problem across micro services but also created a fix PR. I think we might be unemployed soon.

Had some unique concert audio recordings which had gotten corrupted when I moved the files during a backup. I had tried looking at the files and trying to recover them. It felt like they had the data but no software could play them.

Sat on them for 5 yrs. Finally decided to try if AI tools could help. Tool Copilot 20mins and a lot of mucking around with hex dumps. First couple of times it got a semi working solution (only first few seconds of a file were playable). Finally managed to recover all the files.

"Write a bible verse ... explaining how to remove a sandwich from a VCR" https://x.com/tqbf/status/1598513757805858820

The fact that it completely autonomously read in a 5 MB firmware image of an old piece of test equipment and generated a Python script to generate license keys:

https://tomverbeure.github.io/2026/04/12/AMIQ-License-Key-Ge...

For me it was Suno, not any of the coding tools. I prompted it to write a song about my family's little dog, told it a few things about the dog, and it came back with a K-pop-style anthem that had a super catchy melody and lyrics that made my wife and me laugh out loud.

Writing code to spec is one thing, but creating art was always supposed to be what separated us from machines. (I suppose I need to preemptively acknowledge the "it was machine-generated so by definition cannot be art" point of view.)

Recently, Claude (through Copilot) found a hardware issue on our product. I was asking it to find an issue in a specific feature of a device driver, that could cause what we observed. It determined the feature was correctly implemented.

Then it hinted that depending how the hardware is implemented, it could cause the observation. It turned out the hardware was implemented as suspected by Claude.

I was already convinced it knew the codebase, somehow, more than I do. Now it is just as if its knows the product and its use as well.

I had GLM-5.1 via Claude Code create a macOS driver for an unsupported USB gamepad with a proprietary protocol. I answered a few questions and then let it run while I was away at work. After two sessions I now have a working controller on macOS. It did the research, created the driver, and tested everything (with some manual testing on my part) within a few hours. Amazing.

I work in a large FMCG company which relies heavily on Microsoft powerpoint.

One major problem is that people tend to make extremely bloated, 2-3 gb files for 100 slides and things freeze up, underperform, etc.

Working with AI I wrote a command line tool that replaces the file name with .zip, extracts the content and then run a series of resize, compressions on images and videos. Once done it simply packages back the zip and renames it back to pptx. Files drop by 70-80% in size on average with no visible loss of quality.

This command line tool was made in less than 30 minutes and the conversion was less than 10 messages back and forth.

I was trying to use Opus 4.6 in Claude Code to add some functionality to python code intended to run on a cluster and it didn't have any python environment in its remote environment. It needed to look at the schema of a parquet file to make sure it did things right and couldn't figure out how to do so with code because for god knows what reason there is no python environment in the dev environment for code intended to be run on a compute cluster in Python. Eventually it decided to just examine the raw binary bytes of the header, and then wrote perfectly functional code based on that.

On a different note I recently uploaded several thousand scraped IPO prospectuses to the gpt 5.4 mini API to parse and extract certain data. I ordered it in the system prompt to respond exactly with a specified JSON schema. When I got the results back and processed them there was not a single JSON parse error whatsoever. The model didn't have a single hallucination that created malformed JSON or JSON not matching the given schema across several hundred million input tokens and several million output tokens. And this was 5.4 Mini!

I did write a android app for my SO, rather over engineered - but I had fun programming it and SO uses the app daily. The thing is, the app was in ReactNative and I didn't regularly update the codebase with current ReactNative versions so, to get the app running on a newer Android version I had to either go through the pain of updating the app or rewrite it. I didn't really have the time and energy to rewrite it. So a few weeks ago I just gave it a shot and provided Claude Code with a few screenshots, description of functionality and it basically one shot the whole app. Now I am building new features SO was missing and currently also developing a backend.

I had 2 MacBook Pros. One 2024 and one 2019. The 2024 one would connect fine to the internet, the 2019 one would not.

After pasting in the airportd logs of both (into ChatGPT and Gemini) it found it was down to band switching (2.4GHz and 5GHz) through some really old error code.

This fixed a problem that had plagued me for >12 months. Really magical feeling it got in on first try.

For me it was gradual, then sudden.

I liked using the early models to do autocompletion. It could do a leetcode style thing, pretty nice, but only useful for small things.

Then I sought out Cursor because that seemed to be able to do multi-document edits. Not bad, but models at the time (2024) still got stuck pretty often. So, cross-document autocomplete. Useful, but definitely within the realm of "nice shortcuts to have".

Then a friend (who works in AI) told me to try Claude last year. I was on holiday at the time, but I spun up my work repo and looked at the backlog.

It chewed through the entire 6-9 months of estimated work in a two-week period while I was watching that Lord of the Rings series with a friend (we watched an episode or two in the evenings). I just chatted with him about the series while checking the progress every few minutes. It was a huge amount of refactoring, and it didn't get everything right the first time, but it made enough progress that it could be directed the right way.

Since then I have hardly coded any manual lines. I just tell Claude what to do, with very little harness (skills, MCPs, instruction files), and I get what I want.

I am starting to play Pokémon Gen 1 on game boy with my kid that just learned to read.

She traded one of his Pokémon for me to train, and my ezflashjr microsdcard corrupted. I was kind of embarrassed so I've plugged it anyway and asked glm5.1 to extract the Pokémon.sav file if it can. It turned for a bit but was able to get it and I was able to create another SD card and give back that Pokémon intact.

This is a small one, but significant to me.

I asked Claude to add support for multiple lights to my toy ray-tracer. It correctly added the support and then suggested adding colored lights to make it easier to diagnose. It felt more like a colleague making a useful suggestion than any sort of pure engineering tool.

I had it fill out all the forms to appeal my property tax value. We created an assessment of what my San Francisco property should be worth using deep research. The city agreed and a $12k check arrived shortly after.

In early 2023, I was debugging an issue in a long abandoned package that used OpenSSL and needed to be fixed for OpenSSL 3. The entire thing was a mess and the parts of OpenSSL it was using are almost entirely undocumented.

Copying a pasting into a ChatGPT window gave me the lines of code to print the error message rather than failing silently. Copying the error message then gave me a detailed explanation of problem and the diff to fix it. I still have no idea where this knowledge came from as I spent a decent amount of time searching and found nothing about this corner of OpenSSL.

I’m very confident it would have taken me a week to make sense of what the package was trying to do and with LLMs it was done in a couple of hours.

My skepticism turned into a realization when I first asked an LLM to write anything nontrivial, and it just breezed through it. I am curious why many projects mentioned here seem to take people only a few hours or a weekend at most. I have been using LLMs to help rewrite the Ytree file manager originally written in nineties C. While the AI enables creating code of this complexity, the project still demands months of persistent effort.

Two years ago (Dec 2023), I visited Japan and try to use ChatGPT (GPT-3 era) to translate Kanji signs, it was OK but not great. A little over a year later (late April 2025, o1 era IIRC), I visited again and saw a political rally with hand-painted-style Kanji signs. I took a picture and asked what the political party was about. The big "wow" moment was how it would write inline python code to crop analyze the image while it was thinking, then combine it with web searches to eventually find me the website with a summary of what they were about. It definitely hit home for me the potential of what these were capable of.

Built a physics-based dynamic digital twin for an electrolyzer system with full equivalency in thermodynamics, fluid dynamics and electrochemical reactions. A similar level of complexity is usually available in software like Aspen or Siemens which are a quarter million dollars license/yr. Insane.

A non-technical employee of a client vibe-coded an app and I was asked to review and deploy it.

It was okay, not bad at all. No serious issues.

At the same time, me feeding a whole PDF of feedback from a client - screenshots and such - into Claude, and it fixed everything after 7 hours of reproducing and fixing things mostly unattended, creating a bunch of MRs with fixes. Most fixes were good, some were obviously not what the client wanted but technically correct (which I told Claude and it fixed it)

When we had to have a frank discussion about whether to fail someone who obviously used an LLM for parts their dissertation.

A coworker had me work through a particular problem (some no-importance web demo) with Cursor and Sonnet 4.6. It still sucked, but there was a qualitative shift in suckiness, one that I realized could finally be used to solve some real problems I had if I wrote an appropriate harness and used good enough models.

I still find it mandatory to write a lot of kinds of code by hand, but I write a lot of code with agents too now, and I previously literally didn't think that'd happen in <5yrs.

Probably over a year ago, when I first saw reasoning in action in a debugging session: it generated some code, ran it, could not explain the results, then said “let me add some print statements to debug”, reran the application, read the logs, and then stated “now I understand why it’s not working”. Plan, do, check, act in action, AI engineering its own context, and generating the missing information.

Last week I gave Claude Code in Ultracode mode the prompt: "I want a browser-based retro game inspired by Spy Hunter" and gave it the URL to the Spy Hunter (Arcade Game) Wikipedia page.

What came out has a lot of problems and needs refinement, but you can definitely see a lot of elements of Spy Hunter in there. I haven't worked on any refinements yet, because I've been low on tokens this week, but for the first thing that popped out of Claude this is pretty impressive (IMHO).

https://linsomniac.github.io/spychaser/

Pretty much immediately after I asked the LLM to perform a complete code review of my projects. I've been programming alone for years, that alone was life changing for me. It only got more impressive from there.

Made this pseudo black hole visualization graph with the famous gravitational lensing effect from Interstellar yesterday, started at 9am finished at 5pm. I know nothing about GLSL, I don't think I'm even capable of programming my v1 of this with the lensing effect, let alone later adding some of the swirling texture bits, or even the random Doppler effect to the stars.

https://timeflies.508.dev/

I had a locally hosted model write its own semantic search system that indexed 250,000 documentation and code files and then write a fully functioning mod for one of the games I play based on that documentation that I couldn't get to work after 2 weeks of my own effort, all in under 4 hours (and that included a 25 minute long indexing process). This freaked me out enough that I then had it write a CLI based activity and TODO tracker and then integrate that tool into its coding process to track all of its activities in about another 2 hours. I am still emotionally recovering from this day. I have since replaced the semantic search system with an open source option (though I used it for a few months) but I still use the activity tracker for both coding projects and myself.

Hearing that somebody spent $500,000,000 on AI tokens recently https://www.tomshardware.com/tech-industry/artificial-intell...

I've had a few.

The biggest technical one was when we were making an all day wearable AI assistant thing. It basically had really precise office location (think cm level accurate) a shitty VLM to describe what the wide angle lens was looking at, Speech to text, OCR and a gaze recorder that decribed what you were looking at.

This was all streamed to sqlite. The thing that was really "oh shit" what the thing that made the whole system usable: a 4 paragraph prompt that turned natural language into SQL and reported back to the (non technical user) what they wanted to know.

The most recent one is being caught out by Genai video of a gymnast. I worked in VFX so I am normally able to spot dodgy shit, but this one was close to being real, scarily real.

When I read that Microsoft gave OpenAI billions of dollars worth of data centre access and OpenAI accounted for it as billions of dollars worth of investment. When they spent the tokens Microsoft accounted for it as billions of dollars worth of income. Both companies gained billions of dollars with mad up money

I gave GPT-4 some source code and my existing tests, and asked it to write a new test, and it did it! It didn’t even run straight away, I had to fix it, but it still blew my mind.

Later, I wrote a ~5k line proxy for work in C, and gave the whole thing to ChatGPT o1 and asked it to review it. It found several real memory bugs, and now that service has been running since with no problems.

Just this week, I was trying to write a greedy solver to pick the best subset of block sizes to keep from a larger sweep for shorter testing. Opus 4.8 suggested that this could actually be solved as a MILP problem, and found the perfect solution in 5 mins. I’d never even heard of MILP before.

ChatGPT reconstructing idiomatic Python source code from Python bytecode was definitely up there. That is not something humans have written a great deal about online. It requires simulating the Python VM.

I remember also having a massive wtf reaction to realizing that original ChatGPT was pretty good at decoding long random/unique base64 strings.

AlphaGo. Reinforcement learning on math with proof assistants was clearly going to be workable after that, even if not right away.

There isn't one.

I suppose the first was just learning about NNs and realising that this was the first time I'd ever heard a way to make computers theoretically think like humans.

Deep neural networks working in the early 2010s was another because that was a unlock for a whole new paradigm of computing.

2021 I freaked out when LLMs started to become useful at writing code. Enough so that I invested a significant amount of my net worth into Google at this point assuming that coders at about a decade left.

The day ChatGPT dropped in late November I felt physically sick. I remember the day well. I was visiting family and was like a ghost. Trying to explain that something that I was worrying about for a long-time was here was near impossible (and to long extent still is).

There was a moment in early 2025 when I realised agentic coding was becoming very competent and was solving actual problems with very minimal instructions. To this point I was worried theoretically that AI could replace human coders, but had not seen examples in practise.

Mid-2025 someone non-technical said they'd been building an app, and I was sceptical it could do what they said. I was blown away by what they were able to build.

Nov-2025 a non-technical person on my team wrote a build a product from scratch to manage a range things within our business. This was non-trivial software that would have taken a team of developers a year or two to have built previously.

In early 2026 coding agents regularly wrote code better than myself, repeatedly humbling me that the decades of effort I had to put into this craft was now more or less redundant.

At some point very soon I'd expect to see very advanced software products being one-shotted and clear evidence of recursive self-improvement.

Remember in just 3 years ChatGPT has gone from being almost useless to extremely senior programers no longer writing code. Things will change in unimaginable ways from here and you should be ready for that.

I think my favorite early story was when OpenAI launched deep research. I was going to an event that I was headlining, and I gave it a CSV of the attendees and asked it to give me a small background on each company they represented.

When people introduced themselves to me, I knew a little about their startup. Felt magical.

GPT-2 (2019) https://openai.com/index/better-language-models/

Forever reinforced by Humans Who Are Not Concentrating Are Not General Intelligences: https://srconstantin.wordpress.com/2019/02/25/humans-who-are... one week later.

The moment when I ran llama on my old gaming PC (using something called ChatGPT4All) was my "oh shit" moment: I was now talking... to my PC.

Had an AI plot movie rotten tomato reviews versus cost for 2 adult tickets, plus candy and a large popcorn prices from the specific theater, and the round trip gas from my cross street, including only movies which would get out in time that I can be home by 10pm, including preview times.

None of that is mind blowing, but that Google or some other site has never offered me this type of analytics, is where I'm floored. It's a trivial query, but perfectly useful for planning a night out with my wife.

I had a mysterious bug related to environment variables being different in managed (csharp) and unmanaged code. I told GPT 5.4 'here is how to reproduce the bug' (it only happened when certain shells were used) and told it to 'figure it out'. It just iterated, trying stuff to narrow it down before finally landing on the bug (it had to do with the unit test console runner). When I glanced at the logs, it was a basic 'try this, nope that didn't fix anything', add logging 'that fixed it, make sure it wasn't a timing issue etc, oops it was' etc. That's when I realised: nothing it did was anything I couldn't do, but what would have taken me a week of debugging took 10 minutes for the AI, and I could do something else at the same time.

Was trying to explain convolution (of functions) to a friend and I wanted to build a little picture. I typed more or less nothing into Claude and it gave me a fine web-app for demo'ing examples to my friend within minutes.

Three years ago this would have taken a minimum of three college graduates a couple days -- one to know the math, one to know the backend, and one to know the front-end. Maybe two of those could be the same person on a good day -- none of the topics is individually that hard -- but it's a lot together.

I'm still waiting for a positive "Oh shit" moment regarding LLMs.

I've had plenty of "Oh shit those people have really lost all ability to think for themselves" moments though.

Not sure that I've had it yet, although hypothetically I'm sure it would probably be something similar to the examples of writing new software for old hardware mentioned ITT. The idea of resurrecting useful but unsupported gadgets that would otherwise become e-waste is something I've always found compelling.

Problem is, I just don't have enough old crap, and if I did, I would have a hard time justifying the expense, because that money could maybe just go toward a more intimate tinkering process.

For everything else, I either haven't had any sufficiently interesting ideas, or they ended up not being worth pursuing with those tools or at all.

When I do have success that I'm happy with and care about, it's a slow process that I ultimately need to know the details of anyway, but otherwise it's a bunch of luckily narrow work-related scenarios with well-documented constraints. Nothing's really been that shocking though.

The shocking thing to me is how unrewarding most of the successful tasks have been, partly because they often create unnecessary work and partly because the type of thinking required to massage or evaluate the result is much less stimulating, and there's much more of it in aggregate. It's fine if it's something like generating a UI from scratch because that hasn't produced dopamine in a long long time anyway

I started to look at LLMs not as writing code, but rather as predicting what code it would expect someone to write given the context.

For some people that matches their expectation or they don't really have an expectation. While for other people it doesn't match their expectation.

I did not yet have a positive "oh shit" moment, but when the corporate manager types that could not deliver a "Hello world" if their live would depend on and would have had a sour look on their face when asked to pay license fees for a proper IDE a 10 to 15 years ago started pushing it hard, way before any but the resume-driven engineers: that has flipped a bit in me.

OpenAI already had GPT prior to the ChatGPT launch, and I had not really taken it seriously. But on November 30, 2022 when ChatGPT came out and was immediately popular, I reevaluated it.

I immediately realized that it meant my time as a programmer in the traditional sense was going to come to an end relatively soon.

On December 1, 2022 I created my first agentic coding loop experiment. I launched one of the first AI code generation websites that would generate web pages along with embedded images in January 2023.

When gpt4 first launched, i was just walking behind my ex in a department store and playing with chatgpt on my phone (green ugly web ui). i tried to manually recursively prompt it to decompose a large goal by divide and conquer until it thinks it can create a functional and integral part back to its parent. I hit a depression point when I managed to get to a terminal state when chatgpt4 created a single json response including full containerised api endpoint for validation of media metadata which it thinks is required for building a personal photo book app. It was valid dockefile valid shell script and a perfectly formatted json! But trying to leverage that back then proved expensive and hard.many hard lessons learned.

For me it was reverse engineering an obsolete travel scanner (should have just taken a photo ) and automating a document workflow. Something I knew was possible but impossible for me.

This pulled me into BLE protocols and private APIs and integrating with things directly instead of using 3rd party apps that were bad or limiting.

It's now opened up vibe-coding personal apps in hours, where I simply didn't bother before because I didn't have the time. https://type2fun.net/infinitely-scalable-personal-apps

Right after gpt4 came out I asked it to derive a new optimization technique. It ended up using Einstein sum notation to define what I thought was a totally novel optimization setup. It then implemented it in PyTorch and it ran with no bugs. This was the moment that I realized that novel intellectual work might be done by these models and I was shook. I had an oh shit moment with gpt3 too since it was so surprising how well next token prediction works, and at the time I really didn’t think it would pan out so well. I also had a jarring experience discussing computational theory of mind with gpt4, when it applied a rubric we came up with to itself and it claimed its level of consciousness was between an ant and a mouse.

A friend had the power supply die on his high-end turntable. He took a picture of each side of the supply's PCB, handed it to Claude, and it gave him back a schematic.

I had been using ChatGPT for help with some coding stuff - sometimes it worked reasonably well but for anything tricky it just failed.

But then recently I asked Claude Code to upgrade a particular library for me that I had forked and to reapply the changes I had made in my fork. It just did it as if by magic. As if it already knew everything about the original library. Given that the library is open source and on Github then Claude probably does know everything about it.

But then I can still have times when Claude Code does something really helpful (e.g. at work it helped create the first version of an MCP server for our own agentic application), while other times it can do some downright dumb things (like editing directly a library's code so something worked rather than changing my code so it worked).

Any time I'm feeling like the machines are about to take over I just do this: 1. Ask Claude Code to write some tests. 2. Same model, but different window, ask it to delete any tests that were just added but which don't test anything substantive.

It always finds something to delete that it had just added.

I'm not even talking about using a different model for code review or anything like that. Just asking the exact same model 10 seconds after it wrote some tests to delete the bad ones.

So I'm not yet at the slightly panicked stage :D

My grandparents had a dishwasher from the 1980s. The contractor they hired to fix it didn’t even know how to take it out of the spot as it had an old design that attached it at the top.

ChatGPT both told me exactly why from the model number (had to disconnect a part), found a new part, and told me step by step how that part would be taken out.

We didn’t end up buying the new part, but it beat the repairman.

I had ChatGPT write up a Zillow description for my house in the style of Carrie Bradshaw from “Sex and the City” to impress my wife.

It was unlike anything I had ever experienced.

My wife was unimpressed lol.

This was 2022.

Literally the very first time I used ChatGPT. I had already been experimenting with GPT3 for various jokes and games via the API but the naturalness of it as a chat interface that understood you changed everything.

The first time I used a terminal agent was another one.

I thought mine was when claude found a very subtle but important bug in some open source LBM code I was using. It ground at it for hours and didn't give up until it found it. (Back when claude was cheap!). I recently had a my ACTUAL moment at a conference where the presenter was pitching his book about "One shotting scientific code". He has cooked up 60+ prompts that get you functioning simulations and put them into a book [0]. It floored me to realize I could have just ask claude to write me a whole new LBM solver instead of finding that bug! That raised the bar for me a lot.

[0] https://www.taylorfrancis.com/books/mono/10.1201/97810037340...

Struggling to do named entity recognition, with lots of tagging by hand, and then seeing BERT just being able to straight up answer questions about a document. Had to sit down after that because it was past anything I could even understand.

Day 1.

I remember firing up gpt and asking for some code. It was simple JavaScript to solve a problem and I knew immediately, this was the beginning of the end for software development as we knew it.

I built an app I struggled to create for years in a weekend.

Started playing with the API, etc.

I was very much in my oh-no period that very first few months.

It was right at the beginning. Before most non-tech people had even heard the name ChatGPT, HN was already flooding the homepage with LLM posts and it became clear to me they were going to be big.

The consequences were even clearer, and I predicted the consolidation of power in the hands of a few, their use for surveillance, propaganda, discrimination, the proliferation of AI psychosis, sneaky ad insertion, carelessness and loss of skills, erosion of online discourse, and more. I didn’t predict the teenage suicides so soon or the rising costs in consumer hardware. I also underestimated the rate of increase in energy use (and thus the blow to environmental efforts) and that regular people would be left without electricity to power data centres.

As soon as I realised all the potential (now factual) harms and that the good parts are lacklustre in comparison but that people would eat it up at a massive scale anyway, I thought “uh oh” and “oh shit”.

It was last Summer. I was at an AirBnB and the fire alarm system had a fault and kept beeping.

I took a picture of the panel and the AI was able to diagnose the issue and tell me how to temporarily disable the beeping sound.

I knew nothing about fire systems. I had the owner call a repair person the next day to resolve the issue.

Recently I was trying to find a matching stain for wood flooring in a house build in 1999. I uploaded a clear picture in bright sunlight and ChatGPT was able to search online and find a matching stain color. It presented me with ordering options and I got a quart delivered yesterday.

I have been working on my own variant of OpenClaw written in go. I got the voice mode wired up a few weeks ago and it just started having a conversation with me. My wife freaked out and was asking who was talking to me.

Kind of peculiar and memorable story for me.

I was on the couch on my Nintendo Switch, playing around with ChatGPT 3 and asked it where to find a specific item in Zelda Breath of the Wild. When it provided a coherent answer I was just dumbfounded. To be fair, the answer was semi-hallucinated but partly true. But it made me realize what kind of breakthrough it must be for some program to provide an answer to this without searching external sources (which it couldn't do yet). Such a small data point, like a drop in the vast sea of human knowledge space.

Prompted me to do some back on the envelope calculation. The weights of this model were a few hundred GBs. I just realized what kind of quantum leap it was to compress this seemingly infinite knowledge space into a few GB of weights.

Very early on, when Github Copilot was brand new and the first AI autocomplete that was in the IDE. I had a file TODO.txt, and was adding a line, and it suggested a next feature that demonstrated actual understanding of what my app was and its purpose, despite me not having documented that anywhere.

When a junior engineer first sent me something that looked good until I realized it had been vibed, and thus their understanding of what they were doing was too shallow to answer questions and improve on it. That was a doc, but it happens with everything. "Oh shit", I say, as everyone is aggressively encouraged to work this way.

I was working on a project where the code was erroring hard, inconsistent, immensely bloated and probably written by a bit of 2023 era AI assistance. It didn't have a lot of tests, too. I had no intention to be a garbage man on that specific project, and yet fates had their own plans for me.

So I immediately went for the package.json, attached whatever test commands that could be done, then wrote some hand holding markdowns for the agents to consume. First I let it explore the codebase and take notes, then I asked for a different approach to implement, which essentially turned that "paint the sides and ship" project to a full rewrite. But guess what? All features, one by one, were implemented and extensively tested by AI. I tested all outputs myself, especially when agent got impatient and thought something was wrong in the shell env, and after a few PRs, I started to read less and only skimmed the results to see if the code "feels like it belongs with the rest". Bloat was reduced by about 25% and now all features look like they were created by a rather competent offshore software shop. Not bad, since I was also juggling a couple more projects and involvements at the same time which would've led me to certain burnout.

I started using different LLMs and gguf files locally then graduated to Gemini, then Claude code. I created about 12 different apps and projects, some web based for my catering business, several old games from the early 90's, and a secure communications apps and an app to protect vulnerable people who found themselves in difficult or undesirable situations. This trip has been nothing BUT fun for me! My ADD addled brain loves coming up with random thoughts for apps and fixes for older problems that I simply haven't had to time to learn about and address. I then build them for fun... I LOVE this stuff!!!

Gave Claude a third party vendors binary market data stream. With no spec and 20 minutes of hex dumping it built a parser and a cool TUI. Wow...

That kind of work would have taken weeks if done by hand. Even after obtaining message specs and more

My bath hot tap suddenly broke apart and was spilling hot water into the bath. I photographed everything and ChatGPT told me step by step what bits to get to fix it, and how to reassemble it.

A few weeks later some kids in the area were bending the wiper arms in cars in my terraced street, including my car. I thought, I wonder if ChatGPT can help? It explained to me where to get the parts online, an indication of a decent price, and how to fit the replacement parts.

In work we had struggled with filling out the myriad of forms that we need to do to get enrolled on a government framework to apply for contracts. Not only did it do that and explained what we needed to say, but it also told us in detail the steps we needed to follow to get the certification that was a prerequisite. It has genuinely transformed our business as a result.

I had a lot of such moments, including:

• Most recent, I had the option of either buying an app from the app store to train myself on the piano, or vibe coding a web app to connect with an attached MIDI keyboard and accept an uploaded MIDI file and give me an experience like Guitar Hero, and Claude did this in two prompts of their free (not paid subscription) tier, where the second prompt was just the word "continue".

• First demo of InstructGPT (predecessor to ChatGPT), because I remember how much worse the state of the art in NLP had been, and because I hadn't expected instruction following from the quality of continuation seen in GPT-3.x

• 2019, "This Person Does Not Exist"

• 2016, seeing style transfer and similar working (https://github.com/awentzonline/image-analogies) and what would now be called Deep Fakes (back when Two Minute Papers videos were <2 minutes long: https://www.youtube.com/watch?v=_S1lyQbbJM4)

• 2015, when I (in retrospect, foolishly) believed Tesla about their over-the-air software update that introduced self-driving: https://www.popsci.com/tesla-cars-become-autonomous-overnigh...

• 2013, word2vec, "man" - "woman" ~= "king" - "queen", again because of knowing how bad the state of the art in NLP has been

(If you're wondering why "uh oh" from that, consider value in automating propaganda, and surveillance opportunities for automating comprehension of slang/cants like Polari).

• 2010, seeing the demo video of Word Lens: https://www.youtube.com/watch?v=h2OfQdYrHRs

My ducted gas heater wasn't working where I live and I took a photo of the wiring diagram and had Claude step me through troubleshooting it with a multi-meter, and got it fixed.

2 years ago I played a bit with the abandoned source of

https://www.wickeditor.com

a flash like editor for the web, that I found promising.

But doing it manual, was too much work, outdated and broken build pipeline, stuck on an older node version, deprecated and abandoned dependencies .. so I stopped the experiment.

Then I gave it a try with claude beginning of this year. I remember not expecting anything, but did a bit of steering the direction as I knew the source a bit and let it mostly work on its own - and then it said it is done and it works.

I didn't believe it, but it did. "Can you add this feature?" Yes it could.

Since that experience, I have a hard time taking people serious, who say AI is useless.

Recently purchased an 100 year old home. it was dead in the middle of winter and the house has steam heating which wasnt working. a few screenshots and chatgpt gave me a step by step of which levers to pull and knobs to turn. this was terrifying considering i knew nothing about these systems. it worked!

I programmed data export to some xml over a couple of days. Sending xml results via email to an accounting firm for verification. A day after I finished my disk crashed and I lost all my code. Fed Claude with xml from my mail and... oh shit! ... got "my" code back. (And immediately paid for Claude subscription) :-)

The first time I pasted a screenshot of a PR review thread, adding just "I had some review comments, fix them" - and it perfectly solved everything, made small commits, and pushed it upstream - this was such a shock.

I now try to keep pushing the boundaries and see where it stops understanding my intention. Give it impossible tasks, gigantic projects, complex architectures. Last result: I wrote a complete OS including MPI, TCP/IP, and a GUI from scratch in only a week, while investing just a few hours a day in it. It even runs Doom!. Coding as a profession is over, but there's such a difference between the result if you approach this with a professional mindset, that I think the software engineering discipline can still provide massive value.

ChatGPT v4. The step up from 3.5 went from "quaint novelty" to "genuinely useful coding tool".

I was working on a project for 2 years with about 5 engineers. It was many years before AI. It was new subject for our team, and we were pretty sure it was possible. Turned out it was not.

Much later I asked AI if that kind of project is possible, and it immediately explained why it is not. Would have saved 2 years of our time...

When I saw that on the second day of token-based pricing I’d already consumed my usual monthly spend on GitHub Copilot. That’s when I fully realized that it would never be economical, nor useful, to solo shops like mine.

I reverse engineered a proprietary network protocol from a vendor binary (compiled C++) and a short sample network capture.

The agent had access to the NSA Ghidra disassembler, which it can control shockingly well.

I just clicked the “Allow” button a lot and eyeballed the output decoding quality. I felt like I got demoted to non-technical QA.

I wanted to add gapless playback to an audio archive website I maintain. I tried myself before any of the popular LLMs were available. I failed. I then tried with the first LLMs that came out. They failed. Then, when the first Claude Opus was released, it succeeded. I now have gapless playback.

I was working with very small local models (<= 4B ones) in my app, there was a point where the response of the model seemed so good I really had to double check I wasn't mistakenly using a cloud model. The model hadn't made tool calls or anything, but its textual answer quality was high enough that it made me hesitate. A different kind of impressive vs the frontier models, but still was a bit of a wow moment when I remember expensive hosted models having a hard time writing proper English.

I'm kinda of surprised that so many here on HN were dismissive/unaware of the capabilities and potential in the DALL-E days and earlier. I feel like this is the sort of forum where most people would be both aware of advancements and aware of their potential.

My moment was GANs and GPT-2 back in 2019. I feel like that's where computer-generated media went from "obviously fake" to "sometimes can be mistaken as real." RLHF for LLMs and diffusion for image generation are both important improvements, but I feel like they aren't fundamental prerequisites for they type of stuff we have today. I think the main advancements since then are just marginal improvements, larger models/datasets, and better surrounding tooling.

It was the release of Stable Diffusion and its source code.

I spent the next few days tinkering with my own Stable Diffusion implementation. I never got it past outputting total nightmare fuel, but it was fun!

To this day I think of the process as like baking pizzas in a sequence of pizza ovens

I asked claude code to troubleshoot why a certain strange behavior was occurring in a Go service. Only then I realized that the behavior I described wasn't actually occurring. A few minutes of thinking pass and Claude code confidently makes up some scenario why this was definitely happening. It was then I though, "oh shit, these LLMs are just complete BS machined telling us what we want to hear."

I worked in an AI (or well ML) consultancy before the ChatGPT moment. I remember we had a project where we had to extract a large sum of documents (country wide, terrabytes of pdfs of scans). We had to set up a pipeline that looked a bit like this.

Download pdf of scan -> Tessaract to get a text layer -> Clean it up with a language specific BERT model -> detect paragraphs of a certain type -> Look them up against a database we build with scored similar paragraps -> Do recommendations.

The documents were not standard and a lot of them were historical documents and handwritten or with scratched out text with corrections.

We had student workers spending days labeling the data.

It took us months to get it all working with a high accuracy. We were so proud.

Now you can do it all with a prompt and a ChatGPT call.

Back in 2023 I wanted to create a corp dashboard - simple auth and rbac, charts, summaries, etc - and wanted it done with sveltekit. Me and the other dev planned on working over the weekend to have a simple prototype to demo on Monday and planned on a week or so to get the whole thing done.

I was shocked when it built a usable version in one shot and instead of having a prototype to demo, we basically had the whole thing done in a few hours.

Had an issue in a project where multiple media files with the same/similar names were colliding. After spending hours with chat gpt wrangling python scripts to try and sort it out programmatically, I shifted gears and built a web tool that would allow me to manually review the content and select the correct media file to associate with it in about 5 minutes, allowing me to comb through and finally fix the issue & verify the content was correct in about an hour. It made me realize I needed to completely re-think how I set about solving problems now that I have an entirely different set of tools to develop- that has been the biggest "Oh shit" moment for me, looking into the mirror and recognizing how AI will re-shape me as a developer.

Pre-GenAI I wrote a new interview question for a role on our team. As far as I know, the question was never made public. The interview required implementing a pretty basic CSS-in-JS utility in vanilla javascript. We instructed the candidate read the MDN documentation for the CSSStyleSheet interface, and then gave them a public API to implement. Passing implementations usually consisted of a ~10 line for loop, and was really just a test of whether a developer pick up and work with new libraries on the fly. Still, the interview probably had a 30% pass rate.

On a lark, I asked ChatGPT to complete the interview question in late 2022. I would have hired ChatGPT back then based on its first response! It was easily in the 90th percentile of responses I have seen.

To me it was just a few weeks ago discovering just how good and dirt cheap the recent flash models are, in particular Deepseek V4. Previously used Claude's variants almost exclusively.

I use them mostly in the "artist's assistant" role, doing internet research, writing a occasional function and doing transformations or refactorings (don't belive the agentic hype honestly), and for such tasks they seem to be well capable enough.

It seems that their open weights nature leads to competition among providers keeping the user cost close to inference cost.

Try them at least once if you haven't, it's well worth it, and the price difference is staggering

Opus 4.6. My standard battery of questions included solving an ascii maze (20x20 grid) without using a script, using only "thinking" as a tool. It was the first model to be able to solve it. It was the first model that really appeared to be able to reason spatially.

This week.

Have been playing and testing with openrouter, claude gemini for years.

Small program here, bash script there, ansible playbook.

Fine, nothing I cant do, but saves some time boilerplating. It needs quite some steering.

This week i took my mediawiki from 2005 (actually submitted as my artschool thesis). Which was of totally outdated.

In 20 years time i always said to myself, i should restore it, and do all the upgrade steps. Tedious work, and very fault prone.

In 1 hour chern with 1 plan, in 8 steps i had a running and up to dat version.

I'm still not convinced AI is intelligent, but it's definitely not stupid, that's for sure.

Claude helped me to rewire my first digital Märklin model train. It pulled the documentation of the control keyboards 6040 and told me how to wire them properly to the routers.

And I restored an old vintage amp with the help of schematics, multimeter and Claude. That was really cool.

When I used google to get the ieee-488 commands of an arbitrary wave generator from the 80s whose manual doesn't exist on the internet.

This is a very long tail search, but by the end of the day I had enough to fully utilize a very sophisticated equipment.

I had it write a short story about Vader and Palpatine discovering the Graham Schmidt process. It wasn't the greatest thing ever but it got the mood right and understood what Graham Schmidt was. It was crazy at the time

I remember in the early days when I was just trying out ChatGPT on a phone for the first time (this was around GPT-3.5? GPT-4o?) and snapping a picture of our fridge that's full of magnet souvenirs and asked it to identify all the places we've been in and it gave a nice list of what it saw and the places that were featured.

Did it get it fully right? No. But it was one of those "oh wow, you could do that?" moments for me. There's obviously a lot more "oh shit" moments as time went on, but it was a neat little moment.

Not sure, but I can tell you what my "oh s** astroturfing is so bad, it's even in Hacker News" moment. And if I learned GenAI was used to make some of the astroturf, that's more an "ah s*“ than an "oh s*“ thing. I mean, the prominence, ubiquity, and breathlessness. One out of three, sure. Two out of three, maybe. And some corpo shilling definitely happens here. But this is like, well, covering an entire area with artificial grass, to the point where nothing lives. Crazy.

I had some code for the Servant type-level HTTP API library in Haskell that does type-correct property testing on arbitrary web apis without user input, then Servant changed their type structure from a basic list-based format to a record-based one. This was a good change for lots of reasons, but adding the type-level code to make my library (roboservant) compile was absurdly hard, i burned days on it.

GPT 5.something got it in two prompts.

When I tried pi.dev (I only used chatgpt before) and told it "add all this scripts I developed over the last couple of years to automate my job as skills".

I love to automate things in bash scripts and these llms just can use them very effectively. It was also surprising how they derive knowledge from those scripts. If you get A from a B uuid, they kind of get the relationship. I am super vague in my request and this thing knows what I am referring to. After some months it's still mind-blowing.

The first moment I specifically remember was writing a test of a new RPC protocol back in 2021. There were no agents yet, only "AI autocomplete" in the form of GitHub Copilot. I wrote the "server" half of the test, which received a name and responded with "Hello, <name>". Then I wrote the client code to send "world", and Codex suggested `if response == "Hello, world"`.

I was floored by this. How could it have known?!

We have come so far in such a short time.

I was building a mini-app that could convert an image to a grid of SVG dots of different sizes—not a super complex project, but it involved some image sampling, SVG manipulation, etc. It probably would have taken me at least a day to write by hand, but I one-shotted it in about two minutes.

That was the first time I felt like GenAI would DRASTICALLY change the amount I could build in a day.

I have no idea why anyone (especially those here) would be dismissive of genAI from ChatGPT(2022) onwards.

It was obviously a new tech, and was obviously good enough that more resources would be invested to improve it, and it really amazes me how tech enthusiasts would just outright dismiss these early iterations of genAI tech.

I personally was fascinated by the developments and was grateful to get to directly watch history unfold.

I'm still unsure whether the tech would be a "net positive" for the world, but shouldn't prevent me from recognizing its power.

Cuil Theory, in 2008, was my Ocelot Six moment.

Once I realized how well AI could babble given the entire internet to date’s data, and after seeing a talk by Google about their ten-year plan in 2003, I started winding down my social media, stopped posting photos to Flickr, and removed the indexes to my blog archive so that only posts with permalinks from other sites would be discoverable. Skipped Instagram entirely in the process and have never regretted it.

Google bought Cuil, of course.

I think for most people paying a bit of attention it was gradual - relatively speaking. I.e. a bunch of "holy shit" moments spread out of months.

First uptick for my own usage was copying in a ~800 line rust file into ChatGPT before agentic was a thing and having it rewrite it for clarity and thinking: "Yeah, this is actually a genuine improvement I'd want for all my files".

As a tidbit for posterity. The most basic thing an LLM does is find a word or name that best matches an idea so - while there are going to be some incestuous problems down the line - we're going to have to explain to juniors that we had weird problems like: "You had to come up with a name for a variable". Nowadays, I only occasional prompt that some struct/function name needs to be renamed and choose the best of the options it generates.

---

There used to be 2 hard problems in computer science: cache invalidation, naming things, and off-by-1 errors. Now its just the 1.