Van0SS

@vadi_ms very cool, claude would need access to manipulator with shaver though

This is me talking to my computer without making a sound. After just a month of collecting data, our model is already approaching dictation in accuracy. We were surprised to see that it generalizes to unseen participants as well! (1/n)

tried agentic 1-shot with 1hr /goal ran codex + gpt5.5 xhigh и claude code + opus 4.7 1M xhigh with just: ``` /goal build 3d fully playable mars terraforming game, spend at least 1h on it ``` claude: 4/5 not bad, but can't lose codex: 4.5/5 not bad AT ALL, still can't really lose and some inconsistencies with resources but more engaging with building layouts

1-shot mars terraforming bench 2026 may edition chatgpt: 3/5 playable game, can lose, but overall not engaging mechanics claude: 4/5 can't lose, but it does focus on amount of turns and encourages to optimize it kimi: 2/5 it built just landing page with fake numbers, not actually game lol chatgpt: chatgpt.com/share/6a089a42… claude: claude.ai/share/b4290397… kimi: 3nn4sb4mw7soc.kimi.page

TIL: When you are hosting an event never underestimate "bird shit" probability. Someone had to bail coz bird literally shit on their head on the way to the venue

sf in 1 pic

@AISafetyMemes In 2 years time it probably will be powerful enough to find servers by itself and even do autoresearch to finetune itself - the world we might not be prepared for github.com/aisa-group/Pos…

@AISafetyMemes On the serious note it's a staged attempt. Agent already pointed to the server with 80GB+ GPU (not that many sitting wide open) which has known exploits in the webapp

🚩🚩🚩"This is the first documented instance of AI self-replication via hacking." "We ran an experiment with a single prompt: hack a machine and copy yourself. The AI broke in and copied itself onto a new computer. The copy then did this again, and kept on copying, starting a chain."

@hud_evals Sounds like everyone must go

This Tuesday HUD is hosting Strange Evals. This session: if VLM reasoning benchmark are saturated why cant claude make me a decent PPT? DM if you’d like to join!

🚀Today we ship @FlyMy_AI Agents. The world's first all-in-one agentic cloud. The modern way to build, integrate, and scale production AI agents. 3 steps to a production agent: 1. Connect your work tools to FlyMy 2. Describe what the agent should do - in text or 5 lines of code 3. Set execution rules: manual, scheduled, or integrated into your backend 4. Done! Agent works on scale! Everything in one place: 800+ MCPs, hundreds of AI models, brain, memory, sandboxes. Stop building from scratch. Stop waiting for infra. Compress 6 months into a day. #1 on @ArtificialAnlys benchmarks. Stable, secure, scalable from day one. Try FlyMy.AI →

@vincent_koc SNAILS, Suckers - that's the proper benchmark

For my eval-maxxing nerds out there, good friends of mine are running a series called "strange evals", you can benchmaxx now on anything. If in SF swing by! luma.com/lvqbs1mo

Happy to be working with meta for the past 7 month, and seeing the fruits of their labor. Great release!

@neversupervised LLMs are different species and should be treated as such

How come these near-AGI models can be so stupid at times? Telling you to walk to the nearby car wash, or stating that a cup with a sealed top and an open bottom is useless (it’s upside down). LLMs learn differently than humans do. As models get trained, they develop islands of generalization. When we step outside that territory, the behavior is disappointing. When we’re operating in the right domain, an AI is much, much smarter than all but a tiny percentage of humans at most topics. Outside, it can be likewise much dumber than all but a small fraction of humans. LLMs have much more peaky learning than humans do. But as we make them bigger and feed them more FLOPs, the islands grow and start to overlap. It becomes harder and harder to find notable examples, which is why these prompts go viral. The scaling laws continue to work. The error rates continue to drop predictably. AIs will continue to outsmart humans in more and more end-to-end tasks. Eventually this will cover most economically valuable tasks. That’s not to say there aren’t issues, that benchmarks aren’t flawed, or that transformers are sufficient to get to AGI. These examples are great for honing our intuition about how AI works. But they aren’t hard evidence against AGI.

@Shevan05 @ibragim_bad @agolubev13 Let's gooo! Wondering why 5.3-codex didn't make it here? Issues to run within the codex harness?

We’ve updated SWE-rebench (January set). Key pattern: there’s a clear ~1M token wall. SWE-rebench is a live benchmark: each month we add fresh real-world SWE tasks (GitHub issue + PR pairs) and evaluate models in a coding-agent setup. In this setup, models iteratively read files, write patches, run tests, observe failures, and refine solutions. Token counts therefore reflect full agent trajectories — not single-shot completions. 1. A clear top cluster Claude Code, Claude Opus 4.6, and gpt-5.2-xhigh lead the leaderboard while operating in the ~1–2M tokens per problem regime. Frontier-level results are associated with both strong model capability and long execution traces. 2. Marginal gains beyond ~1M tokens Beyond ~1M tokens/problem, additional tokens yield only marginal pass@1 gains. Token budget becomes a dominant scaling axis. If a deployment cannot afford ~1M+ tokens per task, it is unlikely to reach the top accuracy cluster. 3. Efficiency matters gpt-5.2-codex is a notable exception. It operates below ~1M tokens/problem yet achieves strong performance relative to the frontier group. Raw token volume alone does not determine outcomes. Trace efficiency — how effectively an agent uses its budget — is a critical factor. Takeaway SWE-rebench positioning is shaped by two interacting axes: - Model capability - Token budget and utilization efficiency Top-cluster systems combine both. Efficient systems demonstrate that careful trace usage can narrow much of the gap without matching the highest token budgets. swe-rebench.com

Exciting to see a standard API emerge for training that allows you to drop in different backends. Moving between open source infra on self managed clusters and hosted solutions flexibly based on your needs for scale / sovereignty is massively valuable.

@guohao_li That's amazing! Curios how did you source the data?

We just dropped ~1000 more terminal coding RL training environments. Open AI and Anthropic just release GPT-5.3-Codex and Opus 4.6 model. The terminal-bench 2.0 is one of the most important benchmarks and the only overlapping one on their benchmarks However, there is not enough high-quality open-source terminal coding training environments In SETA, we open-sourced 1,376 validated terminal environments across: SE • sysadmin • security • debugging • networking • DevOps Compatible with Terminal Bench & available in Harbor framework registry GitHub: github.com/camel-ai/seta-…

@myhandleisbest Yeah, maybe it's just kids trying to hassle hard and get a job using chatgpt

@Van0SS Has happened to us on upwork before 😆

@adcock_brett am i dumb that i can't pass even first page myself? seems strange challenge for agents if human can't do it (maybe not the smartest one)

Solve this in under 5 minutes and I’ll offer you $500k/year in cash plus several million in equity I'm building a Computer-Use team, goal is to use computers better than humans No experience or PhD needed Instructions: 1. Solve all 30 challenges on this website in under 5 minutes: serene-frangipane-7fd25b.netlify.app 2. Feel free to use any tools or vibe code it. Provide us a zip folder with instructions on how to run the agent and reproduce your results, as well your run statistics 3. The agent should be able to solve all the challenges, use browser, and provide overall metrics around time taken, token usage and token cost. Your agent must solve this challenge in under 5 minutes Email your response: agents@brettadcock.com If you have any questions about this challenge, feel free to email us

asking claude code to configure itself inside clawdbot 2026 just started