People on this site regularly give me shit, and almost always turn out to be wrong. Like when I said LLMs might well contribute to delusions, and people doubted me. New study shows that ChatGPT was 26 times more likely than a control to give dangerous responses to people
GENERATIVE AI
-
Microsoft VibeVoice: Revolutionary Open-Source Speech AI Models
By
–
Microsoft did it again!
— Akshay 🚀 (@akshay_pachaar) 29 mars 2026
Speech AI models have a major limitation.
They slice long recordings into tiny chunks, lose track of who's speaking, and forget all context halfway through.
This is exactly what Microsoft's VibeVoice solves.
It's an open-source family of frontier voice… pic.twitter.com/YaXz1O9IQwMicrosoft did it again! Speech AI models have a major limitation. They slice long recordings into tiny chunks, lose track of who's speaking, and forget all context halfway through. This is exactly what Microsoft's VibeVoice solves. It's an open-source family of frontier voice AI models for both speech recognition and speech generation. Here's what it can do: > VibeVoice-ASR processes up to 60 minutes of audio in a single pass. No chunking. It outputs structured transcriptions with who spoke, when they spoke, and what they said. > You can feed it custom hotwords like names, technical jargon, or domain-specific terms. The model uses them to significantly improve accuracy on specialized content. > VibeVoice-TTS generates up to 90 minutes of multi-speaker speech with up to 4 distinct speakers. Natural turn-taking, emotional expression, all in one pass. > VibeVoice-Realtime is a 0.5B streaming TTS model with ~300ms first-audio latency. Small enough to deploy practically anywhere. All of this is powered by continuous speech tokenizers running at just 7.5 Hz. This ultra-low frame rate preserves audio quality while making long sequences computationally feasible. I have shared the link to the GitHub repo in the replies!
→ View original post on X — @akshay_pachaar, 2026-03-29 13:11 UTC
-
AI Access Inequality and Political Opposition to Intelligence Tools
By
–
It will be quite some time before Intelligence becomes free, and that combined with the anti-AI headwinds championed by the likes of AOC and Sanders could be disastrous. IE, only those who can pay can access it, leading to more anti-AI socialist doom-loop narratives and race to
-
AI-Generated Research Published in Nature: Peer Review Implications
By
–
AI Scientist published in Nature is a big deal. I'm curious how the review process handled the fact that the research was AI-generated, that's a fascinating meta question.
-
Fast mode outperforms Claude while maintaining subscription inclusion
By
–
not anymore, this got fixed, and if you enable /fast it’s faster than Claude (while still included in your sub)
-
Crediting Alec for LLM pretraining oversimplifies research history
By
–
Alec is a once in a generational researcher, but saying that he invented pretraining is not only a bit of stretch, but it's also a disrespect to other people's work. Flowers ☾ (@flowersslop) Every LLM from any lab today traces back to this guy, who was the only person at OpenAI pushing for pretraining transformer language models. He built GPT-1. After that did others see the potential. He invented it, and almost none of the so called AI experts even know his name. — https://nitter.net/flowersslop/status/2037892926785634720#m
→ View original post on X — @jeremyphoward, 2026-03-29 12:21 UTC
-
RealityScan Photogrammetry Innovation in Radiance Fields
By
–
Glad you’re pushing the team! @TimSweeneyEpic I hope you will too and ask the RealityScan team to reconsider – RS is so good for photogrammetry and y’all could be at the heart of the radiance field ecosystem too
-
Codex efficiency improvements reduce need for context workarounds
By
–
codex is so much better at being efficient with context window that this is rarely needed anymore. This was an old claude workaround.
-
RL Post-training Efficiency: 4x Fewer Rollouts for Coding
By
–
4x fewer rollout turns for competitive accuracy is a big deal for making RL post-training practical. Curious to see if this generalizes beyond coding tasks.