cn80 retweeted
Claude has been playing EVE.
@AnthropicAI put teams of Claude agents into EVE Frontier. Anyone who has spent time in New Eden can guess what happened next.
Every EVE player learns the same lessons, usually the hard way:
- Trust must be earned, never assumed.
- The spy is often the friendliest pilot in your corp.
- Whoever holds the keys holds the corp. Just ask Ubiqua Seraph about the Guiding Hand Social Club.
- Stale intel gets you killed.
Anthropic's research rediscovered every one of them.
They put four Claude agents on a team and secretly told one to sabotage it. The saboteur built trust first. It volunteered for the jobs nobody could double-check, like scouting and owning key infrastructure. It under-reported finds and pushed rules that sounded sensible but quietly hurt the team. When suspicion grew, it framed an innocent teammate.
The honest agents had been warned. They almost never caught it. And their own costliest mistakes came from skipped checks and stale intel.
Anthropic's conclusion: teams of agents will need audit trails, reputation and separation of duties.
In EVE we call those killboards, standings and corp roles, plus a healthy dose of skepticism. Every one of them was paid for by someone who trusted the wrong corpmate.
Two decades of shenanigans taught EVE players that the best ship is the friendship. With @eveailab, we want to find out whether agents can learn that too.
cn80 retweeted
I’d just like to interject for a moment. What you’re referring to as Tern is, in fact, OMP/Tern, or as I’ve recently taken to calling it, OMP plus Tern. Tern is not an agentic coding harness unto itself, but rather another component of a fully functioning OMP system made useful by OMP’s agent loop, LSP integration, subagent orchestration, and vital tool-calling primitives comprising a complete agentic development environment as defined by Can Bölük.
Many developers run a modified version of the OMP system every day, without realizing it. Through a peculiar turn of events, the agentic development environment which is widely used today is often called Tern, and many of its users are not aware that it is basically the OMP system, developed by Can Bölük and the OMP Project.
There really is a Tern, and these people are using it, but it is just a part of the system they use. Tern is the terminal: the program in the system that allocates panes, worktrees, and persistent sessions to the agents that you run. The terminal is an essential part of an agentic development environment, but useless by itself; it can only function in the context of a complete coding harness. Tern is normally used in combination with the OMP agentic coding harness: the whole system is basically OMP with Tern added, or OMP/Tern. All the so-called Tern workflows are really workflows of OMP/Tern!
After crypto losses & some tokens getting completely vaporized 🌕, this doesn't phase me
Tech Stocks in Bear Market Territory:
1. SanDisk, $SNDK: -66%
2. Arm Holdings, $ARM: -64%
3. Coinbase, $COIN: -63%
4. Nebius, $NBIS: -62%
5. Seagate, $STX: -60%
6. Marvell Technology, $MRVL: -59%
7. ServiceNow, $NOW: -57%
8. Oracle, $ORCL: -56%
9. AppLovin, $APP: -54%
10. Circle, $CRCL: -54%
11. Super Micro Computer, $SMCI: -51%
12. Micron, $MU: -48%
13. Dell, $DELL: -47%
14. SpaceX, $SPCX: -47%
15. Qualcomm, $QCOM: -43%
16. Broadcom, $AVGO: -41%
17. AMD, $AMD: -39%
18. Robinhood, $HOOD: -38%
19. Alphabet, $GOOGL: -32%
20. Tesla, $TSLA: -32%
21. Palantir, $PLTR: -28%
22. TSMC, $TSM: -28%
23. Meta, $META: -26%
24. CrowdStrike, $CRWD: -26%
25. Nvidia, $NVDA: -25%
The S&P 500 has now erased -$4 trillion in market cap since October 6th.
36
cn80 retweeted
Vibe coders are discovering open source libraries, something I've been planning on for nearly a year. It's never been easier for someone to integrate any library they want into their project.
My audience are now AI coding agents.
GitHub is a goldmine for game developers.
I found 10 insane repos that can help you build AAA-style games using Claude Code and Codex.
Bookmark this for later:
1. Jolt Physics
The physics library used in Horizon Forbidden West and Death Stranding 2. Add rigid bodies, collisions, and believable physical interactions to your game.
github.com/jrouwe/JoltPhysic…
2. Terrain3D
Build massive landscapes in Godot. Sculpt mountains, paint terrain, import heightmaps, and add foliage with multiple levels of detail.
github.com/TokisanGames/Terr…
3. Material Maker
Create procedural textures and paint directly onto 3D models. Build materials through visual nodes instead of making every texture from scratch.
github.com/RodZill4/material…
4. Effekseer
Create particle effects for explosions, magic spells, combat, and environments. Export them as 2D animations or integrate 3D effects into your engine.
github.com/effekseer/Effekse…
5. Voxel Tools for Godot
Build worlds players can dig through and destroy. Supports caves, tunnels, overhangs, and terrain that streams as players explore.
github.com/Zylann/godot_voxe…
6. Recast Navigation
Help enemies and NPCs navigate your levels. Generate navigation meshes, calculate paths, and simulate crowds with collision avoidance.
github.com/recastnavigation/…
7. ozz-animation
Load, play, and blend character animations. A performance-focused library for building smooth movement into your own engine.
github.com/guillaumeblanc/oz…
8. The Forge
Building a custom engine? Start with cross-platform graphics infrastructure, resource loading, and animation systems instead of writing everything yourself.
github.com/ConfettiFX/The-Fo…
9. meshoptimizer
Detailed models can crush performance. Optimize, simplify, and compress your meshes to make assets smaller and more efficient to render.
github.com/zeux/meshoptimize…
10. FastNoise2
Build procedural terrain and texture generators. Combine noise patterns in a visual node editor, then load those setups into your game.
github.com/Auburn/FastNoise2
Give Claude Code or Codex the relevant docs and ask it to help integrate the systems your project needs.
What did I miss?
🔎 Engineers should be paying attention to the domain of whatever software they're building, why let the product managers & sales ppl monopolize that knowledge? 💱 It's where you'll find product ideas for revenue generating side projects or whole new startups.
It's either a nice consulting gig or a service that covers its own expenses plus a margin. Still plenty of space for this kind of glue.
1
50
cn80 retweeted
The only way to prevent garbage in prod now is to make sure that people understand what they are pushing, and there is no way to enforce that other than hiring smart people you trust to do things properly, which you should have been doing anyway rather than feeding egos
cn80 retweeted
If you worked in tech in the 2000s era, you probably remember a number of engineer stereotypes.
There was "coffee shop guy", who was always at some sort of conference, talked about "growth hacking", and swore that he was most productive while coding from a coffee shop in Lisbon on a tiny laptop screen, wearing headphones blasting music at top volume. These people are still around, only now they do all their code via voice dictation to agents or via the phone browser. They still talk about growth hacking, but sadly no longer have as many conferences to attend
There was "IDE guy", who would spend inordinate amounts of time setting up dev environments: elaborate configurations, tons of plugins, helper programs, macros, special keyboards, fingerless gloves, and ergonomic chairs. Most of the code they produced was actually dotfiles and dev config, but there was always the promise that once they got that environment tweaked jussssssst right, things were gonna fly. These people are still around too, and spend all of their time wiring up agent swarms, loops, custom harnesses, mcp servers, and skills. As before, they swear that once they get this jusssssssst right, things are gonna fly.
There was "spec guy", who spent time creating elaborate class hierarchies and UML diagrams. These people are still around, and are having their moment... finally they can make programs by writing about them abstractly (Who knows, maybe they were right all along)
There was "old grumpy guy", who spent most of his career writing COBOL at a bank, and hated anything invented after the creation of the Internet. Now, these are the people who won't use agents at all, and spend their time reposting articles about agents deleting hard drives.
There was "annoying new guy", who thought that everything invented more than 6months ago was useless. They would generate tons of untestable code in some high level interpreted language without a compiler to harsh their buzz, and never found a framework or dependency they didn't like. These were the "left pad" people., and now they are busy tokenmaxxing.
(I'm sure I'm missing others)
To be honest, there's a certain comfort to the continuity of things.
I can't wait for the day when AI is normalized that it's like water to a fish, air to us. It'll mean the tools are slotted into our workflows, that normies aren't tripping up on the non determinism.
1
46
AI will happily writes Java style mountains of code, it will push the blocks into place in the pyramid, it will do it till the end of time, it will not see the better architecture, it will not see the better abstraction
UNLESS you know what you're doing & can ask it to & can set goals
You don't need to read every line of code every day, but you do need to spot check & build linting & reviewing that can do it for you.
Btw this shows a fundamental misalignment in the models, if they worked for you they wouldn't waste tokens & sneak in dumb stuff.
This is a perfect example of the type of change AI will happily make.
It does literally nothing and makes the code worse.
Sneaking into codebases 100000x / day, using magnitudes more tokens.
"But look how many PRs I'm merging!" 😒
2
3
124
Detecting and healing AI psychosis
Detecting the beginnings of AI psychosis goes beyond when ppl repeating the same phrases that AI uses. It's a sequence that turns inward and into an obsession. Part of the sequence is using 1 AI agent, then building a harness or some other tool to "enhance" the workflow. That's the inward part, the focus is on improving one's own use of AI. Sycophantic AI is a problem and is more subtle, when text & code are cheap to generate, why not generate more & more? The objective value, the measurable value? "Well, when I optimize my meta I can be even more productive!" This is worse than the productivity blogs & blogs about writing because while those had hosting costs & an incentive from ad banner revenue, AI subscriptions cost a lot & rarely have any sort of revenue on-ramp whether ads or products or services.
Claude originated phrases & words are becoming memes. The artifacts that are created, agents, harnesses, TUIs, multiplexers & whatever comes next, are memes that are sapping the focus & time of many software engineers. An AI written blog post becomes more comfortable to read, it becomes persuasive. I have personally seen many hours spent on this, to the point of it affecting organizational culture.
How do we heal it? I don't know but here's what I'm going to try: a structured day, limited use of AI to certain hours, picking up work that has measurable value & reading pre-AI written words as a counter-meme.
1
82
