Every AI app worth knowing about. One place. 🤖
Joined September 2024
- Tweets252
- Following32
- Followers12
- Likes2
Wanna try Muse and get free 1 Billion tokens?
How to redeem:
1. Sign up at muse.ai/join
2. Open Settings → Redeem token
3. Enter: Q7HTAG
⏰ Must redeem within 48 hours of joining
Win-win. You get 1B. We get 1B.
Still some spots left!
Pacing isn’t a stop button. It’s hitting the brakes a little so safety can catch up. 🛑
@DarioAmodei says AI is getting better faster than we can check it. His plan has 3 steps — and Anthropic is only doing step 1 on its own right now.
• Step 1: let outside safety people sit inside the company (badge, laptop, the real systems) and say in public if the lab is actually following its own rules
• Step 2: the big AI companies in democracies agree on shared safety rules, so one lab can’t win by cutting corners
• Step 3: try to get the same kind of deal with countries like China — hardest part, and it may not happen
Read his post. The fight isn’t “stop AI” vs “go faster.” It’s whether anyone outside the company can check the work while the model is still being built.
We Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so.
Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, employee-level access to our systems, so that they can verify adherence to our safety measures, report on incidents, and assess models’ alignment during training.
You can read the full post here: darioamodei.com/post/we-must…
AI that does AI research — and took gold. 🏅
@AIatMeta detailed AIRA3: a swarm of long-running research agents that placed 8th / ~4,354 in NVIDIA's Nemotron reasoning Kaggle.
• No central controller — shared forum + files
• Gold-medal ensemble used GPT-5.5 + Claude 4.8
• Internal claim: ~27% GPU kernel latency cut
Science agents just posted a scoreboard.
As a test of our progress to advance the frontier of AI research, in June we entered the next generation of our autonomous AI research system, AIRA₃, in a live Kaggle competition run by NVIDIA to fine-tune a 30B Nemotron model. The challenge was to teach the model to reason better — all competitors had access to the same information and were graded externally on a private test set.
AIRA₃ placed 8th out of ~4,000 teams to win Gold, outperforming human competitors who had access to the same frontier tools.
We believe this is a reliable signal that AIRA₃ can improve a targeted capability of an AI model at a level similar to human experts.
Streaming speech that keeps up with the room. 🎙️
@AIatMeta launched Muse Voice Transcribe — real-time ASR, 20+ speaker diarization, endpointing, and code-switching.
• #1 streaming STT on Artificial Analysis (their claim)
• Live via Meta Model API, Meta AI for Mac, Muse Code
• Adaptive delay for speed/accuracy tradeoffs
Hold Fn. Talk. Keep working.
Introducing Muse Voice Transcribe, the first real-time audio perception model from Meta Superintelligence Labs.
Muse Voice Transcribe delivers real-time streaming ASR, diarization with 20+ speakers, and endpointing. It’s multilingual with seamless code-switching and improves accuracy with language, keyword, and context biasing.
The model ranks first on @ArtificialAnlys streaming speech-to-text and on public diarization benchmarks.
Local AI just lost its spreadsheet phase. 🖥️
@NousResearch updated Hermes Desktop: one click reads your hardware, picks a model, downloads it, and configures the runtime.
• Windows + Linux
• NVIDIA-ready path called out
• Free / open-source setup
No more "what fits my VRAM?" doomscroll.
A stealth coding model just dropped — no model card, just vibes. 🕵️
@opencode launched Omen Alpha exclusively for OpenCode Go: $100 of usage for $10.
• OpenAI-compatible endpoint
• Built for coding agents
• Community guessing Zhipu / Xiaomi unconfirmed
Cheap tokens. Mystery weights. Ship anyway.
Multi-shot AI video that actually holds together. 🎬
@grok shipped the Imagine Video 1.5 agent — powered by Image 2.0 for sharper quality, smarter storytelling, and better continuity across cuts.
• Live on web, iOS, and Android
• Arena Text-to-Video ~#5 early ranking
• Agent layer upgrade (not just a model bump)
Same character. Same world. Next shot.
Move the camera. Keep the world. 🎥
@theworldlabs (Fei-Fei Li’s lab) unveiled Atlas — an omni world model for text, image, video, and 3D with pixel-perfect camera control.
• Up to 1 min @ 1440p from reference views
• Explicit 3D reconstruction from a few photos
• Early access opening for select partners
Spatial intelligence just got a name.
Claude’s coding flagship just leveled up. 🧠
@claudeai shipped Fable 5.1 (GA) and Mythos 5.1 (trusted access) — long-horizon coding, research, and knowledge work.
• Cache reads −75% → ~25–45% cheaper agent runs
• Live on Claude API + AWS / GCP / Azure
• Mythos for vetted cyber + life sciences
If Astra is the computer, Fable is the coworker.
Microsoft just claimed the image Pareto crown. ⚡
@MicrosoftAI launched MAI-Image-2.6 + MAI-Image-2.6-Flash — multi-ref editing, web grounding, and production speed.
• Flash: best price/quality (their claim)
• ~2.8× faster than GPT-Image-2-Medium
• Live in MAI Playground + Microsoft Foundry
Ship pretty. Ship cheap.
