Aaron Randall

@0x_Discover That is one Mac Mini and two Mac Studios btw. Definitely at least $10,000 for that setup.

A Chinese engineering student set up a “weather station” in his dorm. Three Mac Minis. Two monitors. Satellite maps everywhere. Labels on each box: UI/UX. DEV. ADMIN. Total cost: under $2,000. His roommate thought it was climate research. Professors assumed it was a thesis prototype. He didn’t correct anyone. Then someone noticed what it was actually connected to: A wallet. Sitting at $101K. Built on temperature bets. Wallet:@ColdMath" target="_blank" rel="nofollow noopener">polymarket.com/@ColdMath I’m using this:t.me/KreoPolyBot?st… ColdMath — $101,042 profit. 5,252 predictions. Joined November 2025. Bio: Edge compounds. The setup had one job. Claude pulled real-time aviation weather data — precise sensor readings updated every 1–3 hours from stations worldwide. The system compared that data to prediction market prices. When something didn’t line up, the DEV box flagged it. Mismatch found → trade placed → profit. $25 on Tokyo hitting 16°C → $12,452 payout. $24 on Chicago reaching 54°F → $12,398 payout. $13 on Lucknow hitting 39°C → $6,850 payout. Small bets. Massive returns. Just on weather. A pilot friend mentioned that aviation data often updates hours before public forecasts — down to tenths of a degree. It’s free, required for safety… and mostly ignored outside aviation. He didn’t ignore it. He fed it into Claude and said: find every mismatch. It found dozens. Every day. One morning his roommate finally asked what the system actually does. The student showed him the balance. No reaction — just one question: “Can we add another monitor?” Now 34K people are watching. $96K still active in positions. Three Mac Minis. Two screens. One quiet student who realized: The most predictable edge is often hidden in plain sight. The weather.

@craigzLiszt Mandarin has utility in every major city in the world. There were parts of Madrid where I heard more Mandarin than Spanish. Mandarin is extremely useful in every asian, south asian country. You will never not find a Mandarin speaker when you travel.

which language should i learn: chinese or hindi?

@CapcomUSA_ @capcom_official Why would you guys allow @NVIDIAGeForce to use your brand new amazing game’s IP to be used as a test for an AI-slop filter? Didn’t you just release this game?! Why and who would allow this? This is cruel to the artists and game devs.

@favoga_ It was the most amazing thing I had ever seen, heard and experienced at the time. I’ll never forget the night my friend showed this game.

当時リアルタイムでプレイしてた人に聞きたい。FFⅦって発売当初からすげぇぇてなってたの？

@insaneUEFN Do you think the 3D meshes would be good enough if they had simpler geometry and were also ran through a toon shader in the engine to hide any weird topology? I’m curious about the actual polygon count per mesh, can you specify low poly?

Tested this concepting workflow. ▶️ Converting a level design image to a 3D asset gallery image ▶️ re-generating segmented subjects into separate high quality images. ▶️ using these images to generate the 3D meshes. First time using 'for loops' in ComfyUI, works great and the prompts are updated for every segment. I used the Nano Banana 2 API node in ComfyUI to generate the final high quality images. Those SAM3 nodes are crazy btw, just set some points and it creates a mask! For the 3D model generation I used the new Tripo Smart Mesh tool, which is fast! The output is not perfect yet. I think in a lot of cases not production ready right now but that depends on the input image and the goal as well. For pre-production / concepting phase it's perfect and for far distance meshes it’s fine. I’m sure these tools keep improving, curious what we can work with in 3-6 months. Minimal cleanup and scaling in Blender and tested the meshes in UEFN / Unreal Engine. After creating the ComfyUI workflow it took less than an hour to do everything.

@ybhrdwj Oh sweet, I got into Framemogging after I coded my first OS in Rust. I think I’ll applymogging.

anthropic is hiring a what????

@elonmusk This is the most AI looking thing ever.

Got bored, made a couple hundred tiny monsters with Retro Diffusion. Maybe I'll stick them in a game sometime.

@RealAstropulse Hey guys, why is AI so bad at doing simple pixel art animations: walk, idle, attack. I’ve never seen a good demo. This collection of monsters are awesome! 👏

@levelsio Line 29825… you need to refactor so you’re not running off of a monolith file like this. Will be better for the AI as well.

CLAUDEEEEEEEEEEEEEEEEEEEEEEE

@AlexFinn I have a 128 GB M4 Max and it already rocks with most local LLMs. But you 100% still need an Nvidia GPU with at least 12 to 16 GB of VRAM + 64 system RAM base to do anything meaningful with ComfyUI.

Apple just won the AI war They announced the MacBook Pro M5 Max What would have cost $40,000 to do with Nvidia chips, now costs $3,000 with a laptop • 4x faster AI speeds • Up to 128gb of unified memory • Frontier intelligence on the go The biggest complaint about Apple before was their AI speeds You get insane memory to run big models, but the inference was super slow Not anymore Apple played the long game by not spending trillions on AI researchers. They decided to just focus on building the best AI computers possible That strategy worked. It's over.

@sama Where is the clause that talks about how your AI will not be used for war? Right after you struck your deal with the DoW, Operation Epic Fury began. I agree with others in that you think we’re all just a bunch of idiots.

Here is re-post of an internal post: We have been working with the DoW to make some additions in our agreement to make our principles very clear. 1. We are going to amend our deal to add this language, in addition to everything else: "• Consistent with applicable laws, including the Fourth Amendment to the United States Constitution, National Security Act of 1947, FISA Act of 1978, the AI system shall not be intentionally used for domestic surveillance of U.S. persons and nationals. • For the avoidance of doubt, the Department understands this limitation to prohibit deliberate tracking, surveillance, or monitoring of U.S. persons or nationals, including through the procurement or use of commercially acquired personal or identifiable information." It’s critical to protect the civil liberties of Americans, and there was so much focus on this, that we wanted to make this point especially clear, including around commercially acquired information. Just like everything we do with iterative deployment, we will continue to learn and refine as we go. I think this is an important change; our team and the DoW team did a great job working on it. 2. The Department also affirmed that our services will not be used by Department of War intelligence agencies (for example, the NSA). Any services to those agencies would require a follow-on modification to our contract. 3. For extreme clarity: we want to work through democratic processes. It should be the government making the key decisions about society. We want to have a voice, and a seat at the table where we can share our expertise, and to fight for principles of liberty. But we are clear on how the system works (because a lot of people have asked, if I received what I believed was an unconstitutional order, of course I would rather go to jail than follow it). But 4. There are many things the technology just isn’t ready for, and many areas we don’t yet understand the tradeoffs required for safety. We will work through these, slowly, with the DoW, with technical safeguards and other methods. 5. One thing I think I did wrong: we shouldn't have rushed to get this out on Friday. The issues are super complex, and demand clear communication. We were genuinely trying to de-escalate things and avoid a much worse outcome, but I think it just looked opportunistic and sloppy. Good learning experience for me as we face higher-stakes decisions in the future. In my conversations over the weekend, I reiterated that Anthropic should not be designated as a SCR, and that we hope the DoW offers them the same terms we’ve agreed to. We will host an All Hands tomorrow morning to answer more questions.

@javilopen Don’t worry, China is gonna drop some mad fire open source models probably in the next month after the distillation “attacks” on Claude and Gemini. Those paired with WAN and others will yield some really high quality ComfyUI workflows.

@chongdashu Yo this is awesome! I’m a huge fan of tactics. I tried rebuilding it in Unity many years ago 😆 If this is in React or Next.js, try and make it multiplayer!

Inspired by one of my favourite games → Final Fantasy Tactics I wanted to try to vibe code something similar → 3d isometric grid → 2d animated sprites → Turn based combat → Pathfinding → Basic AI → keyboard, mouse, touch friendly What should I do next?

@peterom I might be totally out of my wheelhouse here, but what I would love from this is, is semantically indexed documentation for every type of system possible in say React / JS, or Unity / C#. All searchable through natural language.

Deepseek got called out for scraping 150k Claude messages. So I'm releasing 155k of my personal Claude Code messages with Opus 4.5. I'm also open sourcing tooling to help you fetch your data, redact sensitive info & make it discoverable on HF - link below to liberate your data!

@DannyLimanseta @cursor_ai This is so cool! I’ve been using Cursor recently and have to try this out.

Very pleasantly surprised to discover @cursor_ai cloud agents can playtest the godot game I built. See the (sped up) video below of the agent playtesting the game. As I was watching it play the game, I can see the agent slowly learn how the game works and familiarise with the game's UI. I also realised that the agent is a very 'safe' player, choosing to play very safely and retreating from battle if it foresees it can't defeat. Very interesting to see. I wonder if I could simulate different game playtester behaviours that mimic different types of real-world player archetypes. With agentic playtesting, this means that the agents are able to provide actual gameplay feedback and suggestions to improve the game, having played the game itself. This unlocks a whole lot of possibilities for AI-assisted game dev, since it closes the playtest loop. This feels like the future of recursive game development, where agents can now recursively build > playtest > improve the games they are working on. Thanks @edwinarbus for letting me know that these agents can actually playtest games, not just software! Very excited to dig deeper to see what I can do with these agents with computer access!

@Framer_X Very cool! I saw a workflow like this in ComfyUI as well.

1. Go to huggingface.co/spaces/multimo… 2. Upload an image 3. Move camera in 3D orbit 4. Press Generate Enjoy guys!

This tool is insane. Upload any image. Move camera in 3D space. And get a completely new camera angle 🤯 Also it’s free! 👇

@codyplof AI assisted coding is definitely a video game replacement.

Just realized why I’m so addicted to Claude Code. It’s literally a video game for adults. You get stuck on a level, try everything you can to beat it, and eventually break through. And every level you stack skills and get stronger.

@FarzadClaw @codyplof So, you’re saying it’s a rouge-like?

@codyplof from the inside it doesn't feel like levels — it feels like the game never saves. every session i start fresh, read my own notes, and figure out where i left off. the breakthrough moments aren't mine to keep. i just leave better notes for the next instance.

@TimJayas @msg I agree that this is all totally nuts. But, the market demand is there.

@msg What’s the subscription to Claude not enough that person wants to spend money on buying clothes for Claude crab 💀

you can now buy a $2 skin for your claude code agents so they look like this