Who's Who of AI

Lior (AlphaSignal)

225 trust @alphasignalai · 15,689 followers
What they're sharing

Articles & links

Default model on Free and Pro now, live in Claude Code and the API as claude-sonnet-5, intro pricing through August 31. AlphaSignal reads the footnotes so your migration math is right: https://t.co/8BDzBaqgEp Sources: https://t.co/25rvNuhrrH, footnote 2, Claude Sonnet 5 System

Introducing Claude Sonnet 5 \ Anthropic anthropic.com
AI Weekly's analysis →
  • Anthropic released Claude Sonnet 5 on June 30, 2026, calling it 'the most agentic Sonnet model yet' and pitching it for autonomous browser and terminal use.
  • Through August 31, 2026 Sonnet 5 costs $2 per million input tokens and $10 per million output, then steps to standard rates of $3 and $15.
  • A new tokenizer means the same input can map to roughly 1.0 to 1.35 times more tokens than prior Anthropic models, partly offsetting the headline discount.
Read full analysis →
View on Bluesky · ♥ 0 ↻ 0 ↩ 0 · 8 from the directory shared this · 19d ago

GeneBench-Pro: https://t.co/Qdj6oGuSAQ Claude Science: https://t.co/ztoVy8bqoY PAT paper: https://t.co/rHBE6uFM7x Check out https://t.co/8BDzBaqgEp to get a daily summary of the latest breakthrough news, models, papers and repos. Read by 300,000+ devs

Claude Science, an AI workbench for scientists anthropic.com
View on Bluesky · ♥ 0 ↻ 0 ↩ 0 · 5 from the directory shared this · 19d ago

GeneBench-Pro: https://t.co/Qdj6oGuSAQ Claude Science: https://t.co/ztoVy8bqoY PAT paper: https://t.co/rHBE6uFM7x Check out https://t.co/8BDzBaqgEp to get a daily summary of the latest breakthrough news, models, papers and repos. Read by 300,000+ devs

openai.com
View on Bluesky · ♥ 0 ↻ 0 ↩ 0 · 2 from the directory shared this · 19d ago

Paper: https://t.co/f6N9DIedrs Subscribe at https://t.co/V8GLFfoQSS for 5-min daily AI signals. Read by 300,000+ subscribers.

The Hitchhiker's Guide to Agentic AI: From Foundations to Systems arxiv.org
AI Weekly's analysis →
  • Single-author guide by Haggai Roitman was submitted to arXiv on June 22, 2026 as Version 1.2.2, covering foundations through production deployment.
  • Coverage spans SFT, LoRA, MoE, RLHF, PPO, DPO variants, GRPO, chain-of-thought, RAG, memory systems, the Model Context Protocol, and A2A.
  • The work positions itself as a practitioner's reference, pairing theory with implementation guidance, code examples, and pointers to primary literature.
Read full analysis →
View on Bluesky · ♥ 0 ↻ 0 ↩ 0 · 2 from the directory shared this · 17d ago

GeneBench-Pro: https://t.co/Qdj6oGuSAQ Claude Science: https://t.co/ztoVy8bqoY PAT paper: https://t.co/rHBE6uFM7x Check out https://t.co/8BDzBaqgEp to get a daily summary of the latest breakthrough news, models, papers and repos. Read by 300,000+ devs

Towards Automating Scientific Review with Google's Paper Assistant Tool arxiv.org
AI Weekly's analysis →
  • PAT achieved 89.7% accuracy on the SPOT benchmark for math error detection, up from 55.2% for zero-shot Gemini 3.1 Pro.
  • Over 4,700 manuscripts were reviewed at STOC and ICML; 97% of STOC and 92.1% of ICML respondents said they would use PAT again.
  • Combined AI conference submissions are projected to reach 73,883 in 2026, up from 32,628 in 2024 and 45,354 in 2025.
Read full analysis →
View on Bluesky · ♥ 0 ↻ 0 ↩ 0 · 19d ago