alt Hacker NewsAsk HN: What was your "oh shit" moment with GenAI?
Most of us were amused when DALL-E and its peers went mainstream, and we were quick to point out the obvious flaws.
Then ChatGPT hit the scene and again, many of us dismissed it as a parlor trick that would never amount to much.
Using LLMs for coding initially was a only small step up from basic code completion, and a welcome farewell to Stack Overflow.
I am curious: what was the specific moment that you went from those quaint, dismissive observations to a slightly panicked, "Uh Oh" realization of what these models can do?
Comments
First time using Claude Code I was rather impressed by how quickly I was able to build out a website with Vue and Supabase. Cool. So.......I always wanted to create a iOS app but knew nothing about Objective C or Swift or XCode. "I wonder if Claude Code can build a iOS app for me?".
I went from 0-to-1 and shipped a podcast player into the AppStore in 2 weeks. Not a simulated app on XCode.....literally a fully approved app on the AppStore. Claude Code walked me through installing XCode all the way through to running a final audit on the app so I wouldn't get flagged during review. Mind blown.
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
I wanted to understand the implementation of some numerical algorithms, and the tech reports were not enough.
I cloned the repo of said library, gave it claude and asked it to write a new technical report in math notation, but with annotation with link to the code so that I can pick up the details. It basically one shotted the full report and that helped me re-implement it in "pure python + numpy", "manually".
The announcement of GPT 3, hands down. That's the day that my mind was blown.
Everything after that has been (genuinely significant) incremental improvements. But that announcement was a qualitative step up: we got ""real"" AI that day, something that could pass a Turing test (as common sense envisioned it, without all the caveats added once we learnt of the genuine limitations of LLMs).
I had a C++ actor model which required an Api like the following (std::function):
child->Async(&ChildActor::Method, child, args);
Refactored it to use small buffer optimisation and std::move_only_function)
child<&ChildActor::Method>(args);
And saw a performance jump since no more malloc in std::function.
It also helped me decipher an animation bug in gtlf importer.
Productivity is x4 or higher.
Literally the first time I used ChatGPT, within days of release. It wasn't so much panic as amazement.
It took HN a surprisingly long time to come to terms with the fact that professional SWE as we knew it was coming to an end.
In 2023/2024 we saw a demo of "denial" being a stage of grief live on this site.
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.
The second I realized it removed nearly all blockers as a bootstrapped technical startup founder.
Claude wiped out the need for web and mobile development resources. I bought a Mac-Mini and had iOS apps up and running in days.
I was talking to a software engineer friend for making a demo. This was supposed to be a quick demo and I had sent him 3-4 wireframes. Then I rang and asked causally, "how long will this take?". He said, check back in the afternoon. sure enough, he delivered a full functioning demo in the afternoon. His starting point was my wireframes fed to claude. Wireframes to a working demo in an afternoon. Life has changed, for good or for bad!
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
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.
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.
Genuinely surprised of the breath and level of interaction with this post. It would appear - perhaps we have data to back up? - a distinct _'flavour'_ of post are becomming dominant. A shame.
I was on-boarding to a new company/project about a year ago. Had a bunch of questions about the system architecture and such, but everyone was firing on all cylinders and couldn't spare much time to answer all the questions.
One coworker took some time to ask cursor some questions, and reported that the answer was accurate (I'm guessing he hadn't tried that before).
That was a game changer. I'd been using cursor for simple autocomplete or brainstorming but now I could have it analyze the entire codebase fairly quickly.
FF to now, I've given Claude Code read-only access to GCP logs and database and it's able to debug entire classes of errors and propose solutions.
I watched a friend generate a 10 pages report based on multiple documents, including scientific papers, and it was almost flawless. It would have taken me days.
A milder version of it was Copilot setting up an environment for a Jupyter notebook. What would have been annoying back and forth between googling and docs went like a breeze.
Starting with the days of Siri, i've been evaluating all chatbots of that nature by writing them a meaningless string of text and seeing how they answer. GPT-3 was the first system which instead of refusing to answer or answering meaninglessly has identified that the string of text has no sense.
> that you went from those quaint, dismissive observations to a slightly panicked, "Uh Oh" realization of what these models can do?
Never experienced any kind of panic, only excitement. I told Github Copilot to add documentation to a function and it documented how the code was used even though there was nothing in the function to indicate how it was used. It somehow knew from the code pattern why I was writing that function.
When playing busy Dota 2 (realtime game), it was crashing sometimes. I asked Claude Code any advice (without any hope) and it debugged somehow that I have unstable IP address and a rented VPS server will improve my connection. I could not believe, it worked…
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!
It was about two days after Google released Deep Dream, if you remember, the thing that took a video and filled it with fleeting hallucinations of mostly puppies, fish heads and lizards. I was suddenly struck by the realization "oh shit, this is much more boring and samey than it first appeared to be", and all subsequent gen AI has been similarly underwhelming.
I think it was when the LLM asked me a question at the end of its response. It felt like something other than a machine. Until then the pattern was me asking a question and ChatGPT giving me an answer, with or without hallucination. When it asked me a follow-up question it felt like talking to a being with agency. An entity that has thoughts or ideas or questions of its own.
Lenovo's Fn+Q does not work on Fedora. Gemini resolved this by fixing the Lenovo driver code, recompiling, and deploying it.
Automating my email inbox, I just wanted to split them into folders according to the attachment name but the fields were often incomplete and ended up missing rules, and imap fetch was taking forever and kept failing. In frustration I decided to turn to ChatGPT to split them by messageid which I had never bothered with because the strings were too long to be useful. I initially intended to build a text list of messages and fetch them all one by one but I ended up making chatgpt crush all the instructions into one gigantic python dictionary using the messageid as keys and using it to generate a single pipelined imap call with success flags, dynamic folder naming, cleanup steps the whole works. I was just working on theory of what I knew was possible, and it's the ugliest table you ever saw, but it works and it runs from memory instead of reading and writing values to a temp file and I'd never been able to keep up with that level of nesting before
I couldn’t make a Rockbox (the alternative iPod OS) simulator run on my MacBook M2 no matter how many guides I followed, then I fired up Claude code and by modifying the original source code it made the simulator run and I was able to start developing custom plugins for my iPod. It honestly felt great since I only have basic C knowledge.
Honestly? Probably all the way back to when Nick Walton used the computers at his university to train a custom version of GPT-2 that let players experience a completely open-ended text adventure game in 2019.
As somebody who as a kid had tried feeding IF transcripts into a markov model to generate random rooms for an amateur MUD, this was mind-blowing. It felt like I was playing a version of the “Mind Game” from Ender’s Game by Orson Scott Card.
I'm making a 3D game and I hate flat worlds, a planet is much more elegant, both finite and infinite in gameplay terms since the surface is not expandable, but you can't hit a world border at the same time.
Cartesian coordinates doesn't work well for the player so I wanted a lat/long/altitude grid system.
I could have spent few days walking through stackoverflow and debuging my upcoming flawed implementation.
ChatGPT web version almost one shot the helpers in 2024 and boy, there were a lot of pitfalls.
I was tasked to rewrite an Oracle Apex webapp. 70k lines of PL/SQL. I asked Claude Sonnet 4.6 to read it all and boil it down to markdown file with business requirements. Took about 15-20 minutes, and I got a 700 lines long markdown file to guide me during the rewrite. I've since had great joy using /grill-with-docs!
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.
Code reviews. Code reviews in theory done by humans, but containing copy-pasted inane statements of the obvious. Questions that really did no more than demonstrate a lack of context. Code reviews no longer an educational opportunity for the reviewer, a way they learn and stress their own understanding to create a better product and become a better person, destroyed by the siren song of GenAI producing comments that on the surface seem so helpful and sensible.
"Uh Oh" realization of what these models can do?
The code reviews was just how I first saw it, but the rot goes deeper. The "uh oh" was my realisation of how much these can damage people's professional development. These people will never get better at their job than they are right now.
A lot of what else GenAI does is great, but this is an "Uh oh" indeed.
Maybe when I found out you can use it to run terminal commands, spin up and take down dev environments, and even run other LLMs. Suddenly 90% of the difficulty of onboarding to new repos disappeared overnight and a lot of heavily CLI-based workflows became trivial to automate. Never again do I want to spend hours manually sorting out Python dependencies.
When I read in Oct 2024 how a character.ai chatbot encouraged a child to commit suicide. Uh oh.
When I decided to run codex with Qwen 3.5 27b running on my local machine. Up to that point the most success I have had was with using chat interferences as a Stack Overflow replacement. That was my first real taste of agentic programming, and it was both really useful (genuine productivity gains) and local.
Useful thread. Exciting to see what Will be possible in another few years.
It was when I realized that the collective ethics of humanity was so low that this was actually going to take off.
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) :-)
Mine was testing out the copilot preview in the early days. Testing how well it knew semi obscure public codebases. Started filling out the first few lines and got the entire document word for word in tab complete.
That was the day I realised the plagiarism potential llms has.
The GPT-4 demo. Taking a screenshot of handwritten instructions to build a website, along with a drawing of what the website should look like. Then ChatGPT spit out a working prototype.
Also the live video mode demo later that year.
Then the agentic coding breakthrough in Nov/Dec 2025.
I asked it how to configure haproxy, a tool that I had heard in passing about, and it gave me back exact working configuration syntax for my use case. Today that seems very mundane, but first time that happened, and I didn't have to google, read docs, or worst case sift through code, that blew my mind.
I work with a Go monorepo and set up Bazel for a couple of services that used CGo. It took a while but was painless to set up.
I suggested to a masters' student that a problem we were working on would benefit from analyzing it mathematically. He brought an incorrect solution the next time we met, and on a whim, I asked Gemini to do it. Gemini got it right. I started looking for more ways to use it after that.
It was when I first saw an LLM reliably make tool calls to bash.
When ChatGPT allowed me to calculate stress and load bearing tolerances for a camper based on different materials, suggesting better designs, with the math and sources to back it all up. Then it helped plan and fill out paperwork for a residential solar project, including full code-compliant electrical work, again with sources to verify. Then there was an open source app that wouldn't run on an old version of MacOS due to them not supporting older OSes, and a coding agent backported support for the old OS and got it up and running.
This was my fist ever conversation with Da-Vinci model: https://imgur.com/a/9Cj39MV
When the barriers to actualizing a laundry list of “wouldn’t it be cool to try” dropped was that “oh”. Probably added the expletive when it helped me run headless Blender to rebake texture map and uv unwrap a phone-scanned brown paper grocery bag just so I could find the % surface area covered by ink. It’s more addictive, some might justify as useful, than social media. That is the uh oh.
I tried to get it to generate code to program one of my BitGrid simulators, and it kept producing code that failed, over and over. It was then that I figured out that it can only do CRUD apps and the like, things it's seen over and over in its training data.
It's useless for most of what I want to code.
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.
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.
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.
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.
I was sitting on a cafe listening to a podcast where I heard about a sci-fi author banging out 40+ books per year. How are they doing that?, I thought. Either a team of ghost writers, a boat load of cocaine, or they are using AI.
So I decided to test the frontier of AI, this was back in the early chat GPT era. I downloaded the app and proceeded to go through aln the steps of writing a novel, outline, summary of characters, plot summary, draft chapters, finalised chapters. I had an unedited manuscript by the time I was thinking about my 2nd coffee. It was a terrible novel, but it did have flashes of brilliance that could be harvested and iteratively shaped into something better.
I proved my thesis that AI could mass produce fiction at scale, and If I had a boat load of cocaine the AI and I could probably output 40 books per week.