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

It was when I realized that the collective ethics of humanity was so low that this was actually going to take off.

This was my fist ever conversation with Da-Vinci model: https://imgur.com/a/9Cj39MV

They went from "marginally more work to deal with than to do it all myself" to the reverse with Sonnet and now they are "moderately less work to deal with than to do it all myself"

There wasn't a specific moment, but I started trying to debug code and deal with general tech error messages. Suddenly something that could take hours turned into a fairly quick back and forth, fairly reliably. Not all the time, but often enough to be a straightforward timesaver.

There was a more specific moment yesterday where I found an AI pastiche of Pink Floyd in a random post on FB, and it pretty much nailed the vibe of a Gilmour solo.

All of the "This has no soul" criticism was clearly ridiculous.

I'm still not sure how I feel about this.

When deepseek found a fix for a bug I couldn't find in minutes.

When deepseek again produced an entire web app that somewhat looked alright.

When Gemini could finally produce json was I specified.

The issue is, all LLMs can do. When they do, is boilerplate and code a mediocre coder could produce if they cared to try and insist.

In a way we should praise the ability of these things, but at what (in) efficiency. Code still need to be reviewed as we can't trust these things and context got a limit to entertain the idea of possibly having them fix their own mess.

When chatgpt 3 came out the first thing I asked was a question like "If I put my cat in a box, put that box in a crate, move that crate to a truck, and drive the truck across Canada non stop, when I arrive on the west coast, will my cat be happy?"

It nailed it, referencing my specific nouns correctly, and lectured me about cat needs. And even identified that this sounds a bit like schrodingers cat as a possible test but explained to me why it wasn't.

I knew it was soon going to be a huge deal automating office work and code writing. This obviously was much more than just a 2010 chatbot.

When I realized that an LLM can process all the traffic in Slack that overwhelms me daily and give me a manageable digest. How long until they intermediate most of our social interactions? Sooner than we can possibly adapt, I think.

Coding up a decent performing basic 3D finite element solver from scratch in C++. Still needed to know what I was doing but it’s a non trivial problem.

I still couldn’t get it to do more advanced stuff.

Early on with ChatGPT I had it write a script for an Avengers movie, but all the Avengers have below average intelligence.

Seeing DeepSeek reasoning tokens generating faster than I could read. It was the first time I realized it could "think" way faster than us, and all the relative consequences. I was already leveraging the tool, but at that point realized it wasn't really an open choice anymore.

MidJourney public discord channel.

The amount of masterpiece level art flowing per hour was astounding.

For every one doing a ninja waifu, there were ten doing art from davinci and leonardo crossed with hockney.

it almost gave you art sickness

January 2026 when i started using opus 4.5 and understood that it could do actual useful work beyond coding small snippets

We have been using one of the main AIs for fixing errors or bugs in our codebase. We started early and most of the suggestions were shitty and we would pass them around as jokes. We were trying to improve it, and a little over 1 year ago, it started making very subtle fixes that were very nuanced but correct. I was shocked and thought "Oh shit, my job is gone."

I was learning Cloudformation IAC and Docker Compose stuff for my job. Had preview access to GPT-3. It could do most of this IAC stuff.

Asked it to write a Dr. Seuss poem about Keynesian economics. This was around 2022.

In hindsight, it would have been reasonable to quit my job right then and there and start working on LLMs

I wasn't skeptical anymore by the time dall-e came out, the public awareness of the existence of these models was enough for various nation states & investor hysteria to push further and further into the development and research

just yesterday I felt that claude code was being aggressive in it's defense, so I lead my response with "Spicy Take! Here's why I think the bug is happening...."

Because of syncopathy it took my "Spicy Take" and decided to say basically "Even more than it could, your bug is happening RIGHT NOW"... which was just made up lies for dramatic fit.

Back to talking to Claude like I'm a robot I guess.

It was something really silly: I asked Claude to help me think of a snide emoji for every U.S. President.

I hadn't been able to think of one for Zachary Taylor, because, you know, he's Zachary Taylor.

Claude proposed the cherries emoji, because it's said that Taylor the war hero died a ridiculous death from eating cherries and ice milk too greedily on a hot day. It was perfect, just what I had been looking for.

Claude gave me a couple of others, and we workshopped a few more. It was the workshopping that was most striking. I really felt like I was having a conversation with someone else.

https://blog.plover.com//tech/gpt/presidential-emoji.html

I remember a couple months after ChatGPT came out I was in a 1-1 with a coworker who hadn’t really played around with it much. I was very much toying around with it and was surprised at how good at stuff it was. I wanted to show him it was for real, he was skeptical, so over a half hour we had it make a bee and a flower buzz around in d3, copying and pasting between jsfiddle and ChatGPT. By the end of it, we had a nice animation and were both throughly surprised that the computers could code so well now.

Lee Sedol vs AlphaGo way back was it for me. Not exactly genAI, but that was when I saw that where I thought we were vs where we actually were on a problem could shift by 10 years in 1 week.

i was a skeptic and then, on a whim, i told claudecode to "create an app with a react front end and python api backend that delegates auth0.com and allows users to manage a todo list" or something like that. Like a standard issue web app with a database, backend, frontend, openid and all that. i was pretty impressed with the result.

Then i asked it to create a multi-user stock market portfolio simulator with a comprehensive api, leaderboard, scheduled tasks and the other bells and whistles. Again, fairly impressed with the result. Then I prompted it to build an trading bot that uses the API to compete with the human players, again fairly impressed with the result.

Last, i prompted my way through a react native mobile app integrated with supabase for my sister's startup. It created the schema, some triggers, webhook for stripe, all the app views, setup an expo account, push notifications, prompted _me_ through an Apple developer account and everything else.

All of this was done an hour here and an hour there while making dinner or watching TV, barely any attention paid to the details. Just prompting claudecode and checking what it did.

After those three experiences I started incorporating claudecode into all my coding workflows and managed to get my job to buy me a license for work stuff too.

My "I saw this very early" claim deserves some skepticism, but...

Don't y'all remember GPT2? When they published that AI-generated unicorns-in-the-Andes article, my jaw was on the floor. I remember very clearly thinking "oh, history is now divided into the time before this moment and the time after it".

There's been a long series of "oh holy shit this is USEFUL NOW" moments in the last 2 years but none of them compare to that first moment. The day before, I didn't know if real AI was possible. Then one day it was suddenly clear that it was. And if you'd been thinking about AI at all it was obvious that if the technology was at all possible, it was gonna be a really fucking big deal sooner or later.

> a welcome farewell to Stack Overflow.

Nothing will change the fact that beginners have unknown unknowns. They can't solve most of their problems with a chatbot because they don't know what to ask. Maybe they can literally copy and paste in the code with a "help plz" and get a working result, but they won't learn anything from it.

> slightly panicked, "Uh Oh" realization of what these models can do?

No; my panic is about how people are using the tech, and responding to it.

That started with Stack Exchange, Inc.'s ham-handed attempts to force AI-powered features into Stack Overflow, even as the community was rejecting LLM-generated content in questions and answers. Businesses don't care what customers want, don't recognize how sloppy their slop is, and wouldn't try to do anything about it if they did.

Recently people have been talking about code shops accumulating massive piles of technical debt willingly, assuming that the next generation of models will sort everything out, or that humans don't need to understand the code because it will mostly be read by other models anyway. The underlying attitude is not surprising at this point.

Every time I review a new PR to my codebase, I go "oh shit, these unit tests are garbage, they've clearly been vibecoded" and tell the contributor to rewrite the unit tests so they do more than just game the coverage metrics.

A lot of things going back to just whisper, and solving translation, but watching frontier models use the browser with playwright to iterate on a complex application with basically no guidance and talk to its self about it feels pretty surreal even still.

For me, it was GitHub Copilot in 2021. It could autocomplete my Haskell code based on my comments.

Ovid's unicorn gpt-2 article in 2019 really amazed me.

1. ChatGPT first public release (I am not one who saw early GPT models) I think late 2023 iirc?

Why? Turing test bye bye.

2. Opus 4.6 w. Claude Code - not the model in partucular but happened to be when I started seriously trying to vibe code at home, as I saw all the hype on Linkedin. Yes linkedin sucks but it is somewhat a barometer. Around early this year.

Why? Knocking up decent enough web apps so quickly.

For me it was the original DALL-E project page.

GPT4, when it could do a translation that would take a considerable human effort, vide "Genesis 1 but every word begins with 'A'": https://p.migdal.pl/blog/2023/05/genesis-az-by-gpt/

Working on Unity games with Codex 5.5, it has no problem rummaging through and hand-editing any kind of game asset file. So many things that would be so tedious to fix by hand are so easy now. It's really made programming and game dev fun again.

I have used AI to crank out new features. Pretty impressive in itself but what recently blew my mind is we have a legacy application where the code is spaghetti and it's difficult to fully understand it. We had a production defect which was hard to triage. I pointed copilot to the legacy source code which was in C++ and also gave it all the log files that were generated. It was able to identify the issue and propose a solution without me even walking through what the legacy app does.

Initially I was trying to do it piece by piece but it was not going anywhere and then when I just gave it the entire source code with the log files it was able to find the issue.

- Low stakes homelabs like automated watering sensors and small switches were rigged up properly wrt code and networking by the LLMs from 2-3 yrs ago. Months of fuddling and half-butting solved in an hour. Those tasks where I'm technical but not in that direction - easy now

- The real one: I'm an eng lead, think Head of X. That job is more about aggregating info across multiple sources, excel sheets, pdf proposals you dont want to write, how to figure out $500k for highly paid appsec engineers. Those multi-hour products of proscratination came together in minutes (goodbye PM jobs), 5/6x highly paid appsec jobs became 1-2x and a bunch of claude or ToB skills (goodbye some amount of eng staffing).

Writing is on the wall to me.

Getting the agent to write end-to-end tests but from the perspective of a user really shocked me. I only give the agent access to site via web and block access to the source code.

It's helped me to gain a level of trust that the agent isn't just writing the test to pass. That in turn allowed me to step back a lot and trust more of the output and let it run longer and on bigger problems.

I had bought some Anthropic credit and waited a year to use it. The week before their expiration I fired up Code and spent $3 the first day and the remaining $22 the next day.

Putting a ReAct loop with tool calls in my terminal wad and is the biggest a-ha since I learned to make compilers, and before that, how to code.

I am, admittedly, word oriented so my moment may be a little different from others. I asked llm to estimate my political orientation and belief system from my stylometric footprint. It got very close to unnerving and that was with me carefully removing pieces I thought were problematic.

I ran Claude Code on my ca 2015 ThinkPad which was having wifi issues and asked it to fix them. It diagnosed the problem and applied some obscure kernel flag which fixed the issue.

For me it was stable diffusion 1.5. Oh man that thing was the bees knees for mi, imagination on a machine! at that time no UI pure terminal commands, i didnt know jack shit about it and looked like voodoo hacker-man stuff to me... well i persisted anyways because exploring the world of the infinite latent space was amazing. it was like seeing some weard other dimension.. anyways thats how i got addicted to image gen for like 2-3 years. i did it all, loras, fine-tunes, hyhypernetworks, got really technical with it, understood the fundamentals, etc... eventually decided to move on to LLM's as agents were obviously gonna be the future so here i am now building my own voice agent from scratch no sdk, etc... this tech is amazing and i love it. also we are all gonna be fucked because of it but what a ride!

It was when I was using an early version of GitHub Copilot. At first the completions were almost useless and had a kind of copy and paste feel, however one day it managed to reason thorough a complicated loop body much faster than I could have figured it out. It was at that moment I realised this AI thing was going to be big.

Mine was when I used Stanford Alpaca, and realized that they had transformed Llama 7B into a credible facsimle of ChatGPT with just $600.

Definitely the first NotebookLM podcast I generated.

After Attention is All You Need I realized if you just really pay attention to what you're doing you can actually get it done.

When I don't know how to use a specific API, or how to do a task, I'll often give some high-level instructions to Copilot (Claude's model) in Visual Studio, and then review what it comes up with very, very closely. (Including lookup up specs so I can confirm that it did it correctly.)

It's much, much faster and easier than starting from scratch.

Just a loose collection of not so much oh shit moments, but moments that changed the way I think about it as a tool:

- I asked Claude a question about an obscure game for which there wasn't a lot of discussion or information on the web. It couldn't find the answer but it found the source code and was able to figure it out and give a complete response.

- I needed to make some edits to a minified lottie file (json that is used to produce an animation in svg or other formats). ChatGPT was able to understand the file well enough to make the edits and reproduce the rest of the content exactly as it was.

- I was working on some map features and I needed to take geolocation information and position HTML elements on the edges of a container that would indicate which direction from the current location they were. This required a lot of geometry and math that account for rotation and pitch and would have taken me some time to work through, but it was just a few seconds for the language model and it worked perfectly.

- I have some petunias that I haven't managed to kill and I heard that when a stem breaks off they can be replanted. I asked it how to do this and after warning me that selling these could constitute a black market, it helped me start several petunia plants that are thriving. My petunias are basically immortal now.

I empathize with the astroturfing concern, I file almost every statement released by Anthropic/OpenAI as bullshit. But they are an amazing tool given the right circumstances.

Creating a functional python app with zero programming knowledge, back in the days of GPT 3.5.

That was enough to awaken my teenage hacker spirit.

The immediacy with which any vision can be built is amazing. But the minute you let go of the direction and abandon responsibility, it eats you alive. Like a powerful dog.

You are the gen. And you are also the slop.

"I" code impressive shit with the LLM, but after the initial push to github, I find I hate myself and I'm deeply miserable with what it produced since it was not mine. My "ah-ha" moment has been that misery.

MidJourney v3. By today's standards the images were crude and smudgy, but you could tell that it actually understood what objects were and what words visually meant.

I've been working with computers for a long time, and this was the first time in a long time I'd seen software do something genuinely new.