🔍 Search

Open
🚨 Edison Scientific debuts AI Scientist: finishes 20 studies in 12h

🚨 Edison Scientific debuts AI Scientist: finishes 20 studies in 12h

ChatGPT Query Interrupt, Cursor's new Code Retrieval, ElevenLab's self hosted LLMs, Perplexity's First Research Paper.
Stay updated with today's top AI news, papers, and repos.
|Read on Web
AlphaSignal Logo
Hey James
Welcome to AlphaSignal, the most read source of news by AI engineers and researchers.

Every day, we identify and summarize the top 1% of news, papers, models, and repos, so you're always up to date.

Here's today's roundup:
Summary

Read time: 4 min 29 sec

Top News

▸ Edison Scientific launches Kosmos, an AI system for full scientific discovery

Signals

▸ OpenAI adds interruption in ChatGPT so you update long running queries
▸ Cursor introduces semantic search to agents for higher accuracy in repos
▸ ElevenLabs unveils a unified platform to mix open-source LLMs with their speech stack
▸ Perplexity ships open-source kernels for MoE models in its first research paper
▸ GitHub announces MCP Registry to centralize discovery and installation of MCP servers

MLOps Community

▸ Join free November 18 event to learn how Meta and Microsoft deploy agents

Trending Papers

▸ Tencent and Tsinghua show LLMs can skip next-token prediction
▸ Google DeepMind's new paper reveals the framework behind its IMO win
▸ LMArena launches Arena Expert to evaluate actual expert-level prompts

Top Read

▸ Learn how AI scaling disrupts economic theory
Top News
Edison Scientific unveils Kosmos, an autonomous AI researcher that matches six months human research in one working day
3,932 Likes
Kosmos

Futurehouse just announced the launch of its commercial spinout Edison Scientific, debuting an autonomous AI research system called Kosmos, a tool built to handle full-scale scientific discovery.

Kosmos doesn't chat; it investigates. It reads thousands of papers, writes and executes code, tests hypotheses, and delivers a report you can audit line by line. Beta users say one 12-hour run matches the work of six months in the lab.

The Problem

Scientific research still moves at human speed. Reading literature, analyzing data, and forming hypotheses take time, and even with better compute, humans remain the bottleneck.

Edison Scientific set out to automate the entire research cycle, not just parts of it, while keeping outputs traceable and reproducible.

The Insight

Previous AI scientists, like Robin or Sakana's AI Scientist, could reason or search, but not both coherently. Kosmos fixes this with a structured world model, a shared memory that links every agent's findings together.

It acts as the system's long-term brain, ensuring that every step, from literature review to statistical analysis, aligns with the core research objective.

The Breakthrough
  • Executes 42,000 lines of code and reads 1,500 scientific papers.
  • Coordinates 200 agent rollouts across data and literature tasks.
  • Produces a fully cited scientific report, with every statement traceable to code or source.
  • Completes up to 20 research cycles in about 12 hours.
The Impact
  • Independent experts verified 79.4% of Kosmos' claims as accurate.
  • Reproduced three unpublished or post-cutoff discoveries across fields.
  • Proposed four new findings, including evidence that SOD2 reduces cardiac fibrosis in humans.
How to Use It

You can access Kosmos through Edison Scientific's platform. Upload your dataset, define a research goal, and start a run. Each run costs $200 (200 credits at $1 each), with limited free access for academics.

After about a day, you receive a fully auditable, literature-backed research report showing every data point and code path used to reach its conclusions.

TRY NOW
Signals
1 OpenAI introduces interruption support so you modify deep research queries while preserving context state 4,839 Likes
2 Cursor reports 12.5 percent accuracy gain from semantic code retrieval and trains custom embedding model for it 1,039 Likes
3 ElevenLabs hosts LLMs inside Agents Platform to cut latency and reasoning cost for voice agents 957 Likes
4 Perplexity publishes open-source inter-node kernels that finally make multi-node MoE scaling practical on AWS 822 Likes
5 GitHub publishes MCP Registry that standardizes how you browse, install, and govern MCP servers 738 Likes
Agents in Production - Free Virtual Conference.
November 18 - 3:00 PM CET / 9:00 AM ET / 6:00 AM PT
From Meta and Microsoft to Inworld AI, teams are redefining how agents power search, shopping, and personalization.

Expect technical depth from 25+ talks and a panel on hardening agents for e-commerce - from RL alignment to large-scale reliability.
Trending Papers
Tencent and Tsinghua show LLMs can skip next-token prediction 6,937 Likes
Tencent and Tsinghua propose CALM, which stops predicting one token at a time. It compresses chunks of tokens into continuous vectors, then predicts the next vector. This cuts generation steps and shows you can scale language models without discrete token limits.
Google DeepMind's new paper reveals the framework behind its IMO win 1,535 Likes
This paper introduces IMO-Bench to evaluate IMO-level math reasoning. It shows how to measure math reasoning across answers, proofs, and grading ability, and it provides the exact setup used in the recent IMO-gold model effort.
LMArena launches Arena Expert to evaluate actual expert-level prompts 871 Likes
It introduces a new Expert leaderboard, and also maps all prompts into 23 occupational fields. This shifts LLM evaluation beyond static benchmarks and into expert-grade task difficulty. It lets you see if a model can solve real advanced problems in domains like software engineering, medicine, and law.
Top Read
How AI Scaling Disrupts Economic Theory
alpha_signal_image_1
Harvard computer scientist Boaz Barak published an essay on how rapid AI progress might disrupt traditional economic models.

He uses data from AI capability scaling to test whether current theories of productivity and GDP growth still hold under exponential improvement.

Key takeaways from the post:
  • METR data shows top LLMs double their task-solving horizon every six months.
  • Historical U.S. GDP per-capita grows ~2% annually, doubling every 35 years.
  • Standard models can't handle a labor source that scales both in quantity and quality exponentially.
  • If AI creates virtual workers whose number and skill both rise, GDP could increase sharply.
  • Growth remains capped by the share of work that resists automation even if AI productivity becomes infinite.
  • Real-world frictions like coordination, physical work, and diffusion, limit how fast theory turns into output.

Deterministic AI scaling challenges economic assumptions built for gradual change, forcing a rethink of what "growth" means in an automated economy.
KNOW MORE
At Alpha Signal, our mission is to build a sharp, engaged community focused on AI, machine learning, and cutting-edge language models, helping over 200,000 developers stay informed and ahead. We're passionate about curating the best in AI, from top research and trending technical blogs to expert insights and tailored job opportunities. We keep you connected to the breakthroughs and discussions that matter, so you can stay in the loop without endless searching. We also work closely with partners who value the future of AI, including employers and advertisers who want to reach an audience as passionate about AI as we are.

Our partnerships are based on shared values of ethics, responsibility, and a commitment to building a better world through technology.Privacy is a priority at Alpha Signal. Our Privacy Policy clearly explains how we collect, store, and use your personal and non-personal information. By using our website, you accept these terms, which you can review on our website. This policy applies across all Alpha Signal pages, outlining your rights and how to contact us if you want to adjust the use of your information. We're based in the United States. By using our site, you agree to be governed by U.S. laws.
Looking to promote your company, product, service, or event to 250,000+ AI developers?
WORK WITH US
Unsubscribe" style="font-family:Helvetica, Arial, sans-serif; font-size:14px;line-height:20px; text-decoration:none;color:#DD4825 !important; mso-color-alt:#DD4825; mso-style-textfill-fill-color:#DD4825;" target="_blank"> Unsubscribe

Post a Comment

0 Comments

Users_Online! 🟢

FOUNDER/AUTHOR

FOUNDER/AUTHOR VHAVENDA I.T SOLUTIONS