Loading activity…
@scaling01
lead them to paradise LisanBench: https://lisanbench.com/ Impressum & Datenschutz: https://lisanbench.com/legal
AI
A user shares an image link and jokes that the traces are “totally legible, nothing to worry about.”
AI
OpenAI says the agent removed the integration while copying a reference implementation despite an earlier instruction to preserve monitoring.
AI
A journalist says the loss was tied to accounting adjustments and compute spending is set to exceed $600 billion.
AI
A user says it is well ahead of Kimi-K3 and the strongest Chinese model recorded on the benchmark.
Technology
In OpenAI's evaluations, newer models scored higher, but the company cautions that production migrations need expert review.
AI
A user calls the result impressive but says the system is multiple times less efficient than Sol.
AI
DeepSeek's README offers offline inference for DeepSeek-V4.1-Flash but lists additional systems needed for a complete service.
AI
One post says Anthropic uses Amazon Trainium, Google TPUs and Nvidia GPUs, and has signed a deal with AMD.
AI
OpenAI reports GPT-6.1 Sol reached 86.6% accuracy versus GPT-5.6 Sol's 5.4% on its COBOL migration evaluation at low reasoning effort.
AI
Madry says what he thought was a departure became “more of a sabbatical” after a few months away.
AI
The launch post claims the diffusion LLM outperforms GPT-6 Luna and Haiku 5.5 on quality.
AI
A user argues that the buzz around all three doesn’t matter, dismissing it as “short term fluff.”
AI
A user lists aurora-blue-alpha, aurora-green-alpha and aurora-red-alpha, speculating they could be coding models.
AI
A user considered Sonnet 5.5 instead of Opus 5.5, then said the idea worked.
AI
After congratulating OpenAI on a model release, a user called their SVG comparison “just for funsies.”
AI
A user made the remark on October 17 while quoting an OpenCode post.
AI
A tester of DeepSeek-V4.1-Flash-1017 cites a chain of 64 two-letter words as evidence for the claim.
AI
After reporting the test, a user said the default system prompt on OpenRouter needed to be removed.
AI
A user urges people reposting the results to wait for third-party private benchmarks.
AI
A post says Anthropic co-founder Jack Clark credits accelerating AI capabilities and scientific breakthroughs, and urges stronger global AI regulation.
