> M2.7 is a 230B MoE model that partially trained itself. It scores 56.22% on SWE-Bench Pro and is priced at $0.30 per 1M input tokens. > Music 2.6 brings cover generation and BPM controls. > TTS covers 40 languages with seven emotional registers and real-time low-latency
INNOVATION
-
Spectrum: Open-Source Framework for Multi-Platform AI Agents
By
–
🆕Your agent is ready. The platform shouldn’t be the hard part.
— Futurepedia – Learn to Leverage AI (@futurepedia_io) 22 avril 2026
Meet Spectrum — an open-source framework that brings agents to iMessage, WhatsApp, Telegram, Slack, and SMS/RCS via one unified API.
Bring agents to millions. Natively.
🔗 https://t.co/4jeJ9hZkgJ pic.twitter.com/RFw7vzn2kGYour agent is ready. The platform shouldn’t be the hard part. Meet Spectrum — an open-source framework that brings agents to iMessage, WhatsApp, Telegram, Slack, and SMS/RCS via one unified API. Bring agents to millions. Natively. https://
photon.codes/spectrum -
OpenAI Image Generator Routing Issue Under Investigation
By
–
Someone from OAI has reached out to me and it seems like this image was *NOT* generated with Image 2, but it was routed to a different image generator due to "decision boundary issue." They are looking into this now.
-
Codex Automation Routing: Untapped Leverage in Scheduled Computer Use
By
–
You are using Codex wrong.
— CHOI (@arrakis_ai) 22 avril 2026
Everyone is talking about its coding benchmarks. Nobody is talking about its automation routing. It is severely undervalued.
If you aren't using the scheduled "Computer Use" capability, you are leaving massive leverage on the table.
By tying Codex to… pic.twitter.com/RmovahiNljYou are using Codex wrong. Everyone is talking about its coding benchmarks. Nobody is talking about its automation routing. It is severely undervalued. If you aren't using the scheduled "Computer Use" capability, you are leaving massive leverage on the table. By tying Codex to
-
AI21 Maestro Reaches #1 on BrowseComp-Plus Leaderboard
By
–
We just hit #1 on the @huggingface BrowseComp-Plus leaderboard. Best accuracy: 92.53%. Best recall: 88.79%. Lowest calibration error across all submissions. Built with @AI21Labs Maestro. https://
huggingface.co/spaces/Tevatro
n/BrowseComp-Plus
… -
Computer Vision Transforms Retail with Real-Time Shelf Monitoring
By
–
Most people think AI in retail is about self-checkout, but the biggest impact is happening behind the scenes. Computer vision is now used for shelf monitoring, loss prevention, and safety by tracking inventory, detecting risks, and identifying issues in real time. These systems… pic.twitter.com/GvDwYxrfKc
— Satya Mallick (@LearnOpenCV) 22 avril 2026Most people think AI in retail is about self-checkout, but the biggest impact is happening behind the scenes. Computer vision is now used for shelf monitoring, loss prevention, and safety by tracking inventory, detecting risks, and identifying issues in real time. These systems
-
U.S. Must Re-industrialize to Lead in AI and Future Industries
By
–
Re-industrialise.
— Nina Schick (@NinaDSchick) 22 avril 2026
The U.S. (or West) cannot lead in AI, and the industries of the future (space, energy, advanced manufacturing, semiconductors, biotechnology, robotics) without an industrial base.
Re-industrialise! https://t.co/JzrHqcegSMRe-industrialise. The U.S. (or West) cannot lead in AI, and the industries of the future (space, energy, advanced manufacturing, semiconductors, biotechnology, robotics) without an industrial base. Re-industrialise!
-
GDM Responsible for 75% of New Code at Google
By
–
Translation: GDM is responsible for 75% of all new code at Google.
-
Institutionalization stagnates scientific progress and research
By
–
Hope so. Institutionalization of research has played a major role in in the stagnation of science over the past half a century.