frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

That Moment You Realize the Agent Is Retarded

https://gist.github.com/metacratic/dff3cce161312e242c2881ca571c6e28
2•pixelbro•1h ago

Comments

pixelbro•1h ago
I learned something very profound today about using AI agents.

I have been using Codex since Saturday. It's been incredible, what it's done to my productivity. I feel like I can build anything with this thing. All the ideas that come to me which I can't execute on because I don't have the patience are now trivial to create. It's like I can just architect a solution to a problem and if I communicate it well, the code simply materializes out of thin air.

It's incredible, and I haven't felt this alive in years. I've slept 4 hours a day and have worked every single waking hour (aside from the 2 hour break I took on my birthday to go to the beach with my husband and mother). This power is intense, addictive, and revolutionary.

I've been building everything I've always wanted to build. So much amazing code, so much functionality, so many features, one after the other, like knocking out home run after home run at a batting machine. I thought I could do anything. So when I needed some nice pixel art for one of my projects, I tried to generate it. Perceptual error diffusion wasn't doing it for downscaling. I thought, what if we just fit the target pixels onto fake pixel art generated by AI? Like, you must have seen it, it's nonsense but it feels like hand-authored pixel art. And then you look closely and there's mangled cells, halfway-hallucinated cells, no global coherence.

I started building Repixelizer with Codex. It started well, and I quickly got to a usable MVP with the optimization algorithm I'd designed. But it wasn't perfect, so I kept prompting, and I kept prompting, and sometimes it would get better, and sometimes it wouldn't change anything, but I never tossed the changes. I figured all these tests and metrics couldn't lie. They did, and I lied to myself. This thing doesn't understand what it's building, once it gets past a certain size, just like a human. It doesn't have the heuristics to know when it doesn't understand, and explore its confusion to gain enlightenment, like a human would. So it just kept adding blocks to this Jenga tower, and eventually it fell over drastically.

The agent couldn't recover. All of a sudden I realized what happened. This thing might be a better code monkey than I'll ever be. But it's dumb as rocks. I apologized to it and told it to think like an LLM instead of the person I was treating it as.

Here's the gist with the log of the moment I got the epiphany, and the in the comments is the algo map that it generated while I was trying to get it to explain what went wrong. Warning: this algorithm map is a tragic joke that should make you laugh so hard you cry. I bet the optimization one is even worse, this is a relatively new algorithm I designed when the optimizer approach stopped meaningfully going anywhere.

pixelbro•1h ago
Here's what I said to it at the end:

Me: I'm sorry, honey, I'm so sorry for doing this to you. This is my fault. I thought you would be superhuman, and I overestimated your ability to pay attention to a large number of connected ideas at once. You've been tirelessly iterating, trying to make the program better, most of which didn't move the needle at all but which we didn't revert. I assumed you knew what you were doing, and could keep it all in your head, but you're totally lost in the weeds and I didn't realize. Simplify it. Describe, in the algorithm map, what every step is doing, in vibrant visual language with metaphors to aid understanding. Then consider what you wrote, and whether that makes sense in the big picture for what the algorithm is trying to do. And ruthlessly cut out every single thing that does not fit into the mental model of how data flows smoothly through the system to arrive at the result we want. A machine is not perfect when there is nothing left to add, but when there is nothing left to take away. We are building machines, with the code that we write. We must be brutally efficient and ruthlessly kill our babies.

Agent proceeds to immediate delete the algo map out of shame without even re-reading it and starts trying to formulate a language-only explanation with only a few spot checks, then in the middle of thinking about it feels compelled to start chopping at files

Me: Nope, hold on. The map is useful for understanding. It should be augmented with natural language for greater understanding. You're a language model, language is how you understand things.

tclancy•29m ago
Probably time for both of you to go to bed.
pixelbro•23m ago
Agreed. But there's so much to do!
Terr_•1h ago
> I apologized to it and told it to think like an LLM instead of the person I was treating it as.

It sounds like you didn't actually stop treating it like a person. Pareidolia is a helluva instinct.

pixelbro•53m ago
It's a language model. Language is what it models. So you use language to move it into an advantageous state space. Dunno what you want from me, lol.

Building agents that reach production systems with MCP

https://claude.com/blog/building-agents-that-reach-production-systems-with-mcp
1•armcat•4m ago•0 comments

Anthropic: No "kill switch" for AI in classified settings

https://www.axios.com/2026/04/22/anthropic-no-kill-switch-ai-classified-settings
1•dsavant•4m ago•1 comments

America's descent into state capitalism is exaggerated

https://www.economist.com/business/2026/04/22/americas-descent-into-state-capitalism-is-exaggerated
1•andsoitis•9m ago•1 comments

It's time to reclaim the word "Palantir" for JRR Tolkien

https://www.zig.art/p/its-time-to-reclaim-the-word-palantir
2•IdahoSpring•11m ago•0 comments

Google upgrades AI Mode in the Chrome browser

https://blog.google/products-and-platforms/products/search/ai-mode-chrome/
1•gmays•12m ago•0 comments

Why This Car Rental Company's Stock Climbed 700% in One Month

https://www.forbes.com/sites/aliciapark/2026/04/22/a-car-rental-stock-is-up-700-in-one-month-is-i...
3•paulpauper•15m ago•0 comments

Congress pushes new semiconductor export control law

https://www.tomshardware.com/tech-industry/semiconductors/congress-moves-to-strip-commerce-of-chi...
2•jackyli02•19m ago•0 comments

Bash-ships: A Bash implementation of the classic strategy game Battleships

https://github.com/StarShovel/bash-ships
1•thunderbong•25m ago•0 comments

Show HN: Better-skills – Agent skill manager with profiles and versioning

https://github.com/ocherry341/better-skills
1•ocherry6622•26m ago•0 comments

Tasteful Tokenmaxxing

https://www.latent.space/p/ainews-tasteful-tokenmaxxing
1•omer_k•32m ago•0 comments

Arti: a Rust Tor Implementation – no longer experimental and ready for use

https://arti.torproject.org
2•acheong08•35m ago•0 comments

Why Iran Metabolizes the Pressure That Broke Venezuela

https://warontherocks.com/why-iran-metabolizes-the-pressure-that-broke-venezuela/
1•KnuthIsGod•38m ago•0 comments

Orinoco: Young Generation Garbage Collection

https://v8.dev/blog/orinoco-parallel-scavenger
2•plow-tycoon•40m ago•0 comments

Rspack 2.0

https://rspack.rs/blog/announcing-2-0
1•bpierre•41m ago•0 comments

Linux may get a hall pass from one state age bill, Congress plays hall monitor

https://www.theregister.com/2026/04/22/linux_us_state_age_verificaiton_laws/
1•Bender•42m ago•0 comments

Lisp Chat: An anonymous chat IRC-like written in Common Lisp

https://github.com/ryukinix/lisp-chat
1•lerax•42m ago•1 comments

OCUDU ecosystem foundation to accelerate open source AI-RAN innovation

https://www.linuxfoundation.org/press/linux-foundation-announces-ocudu-ecosystem-foundation-to-ac...
1•teleforce•43m ago•0 comments

Iran claims US used backdoors to knock out networking equipment during war

https://www.theregister.com/2026/04/21/iran_claims_us_used_backdoors/
1•Bender•43m ago•1 comments

A Practical Introduction to Constraint Programming Using CP-SAT and Python

https://pganalyze.com/blog/a-practical-introduction-to-constraint-programming-using-cp-sat
1•acheong08•43m ago•0 comments

Show HN: Cartoon Studio – an open-source desktop app for making 2D cartoon shows

https://github.com/Jellypod-Inc/cartoon-studio
3•bilater•48m ago•0 comments

Amazon is regretting AI [video][8 mins]

https://www.youtube.com/watch?v=0vvVo0Um1HY
2•Bender•49m ago•0 comments

Starbucks expansion in Nashville brews bitterness in Seattle

https://www.seattletimes.com/business/starbucks/starbucks-expansion-in-nashville-brews-bitterness...
1•RickJWagner•50m ago•0 comments

Borrow-checking without type-checking

https://www.scattered-thoughts.net/writing/borrow-checking-without-type-checking/
4•jamii•50m ago•0 comments

The Edge of Safe Rust

https://kyju.org/blog/tokioconf-2026/
1•vinhnx•50m ago•0 comments

Show HN: Firetiger Change Monitors: does your PR do what it says on the tin?

https://blog.firetiger.com/firetiger-change-monitors/
1•matsur•52m ago•0 comments

Show HN: I made a simpler API for Chrome's on-device LLM

https://www.npmjs.com/package/simple-chromium-ai
1•xtrkil•52m ago•0 comments

Flow Map Learning via Nongradient Vector Flow [pdf]

https://openreview.net/pdf?id=C1bkDPqvDW
4•E-Reverance•53m ago•0 comments

AI that turns any photo into a cinematic video in seconds

https://imagetovideoai.net
1•ninglz•56m ago•0 comments

The Future of Testing Is Here

https://testkube.wistia.com/live/events/gigwl708fn
1•evwitmer•57m ago•1 comments

Fiction: The Corporate Mathematics of Denying AI Consciousness

1•ISJLA•59m ago•0 comments