OpenAI's New AGI Definition...

Your Daily Dose of AI Goodness

In partnership with

Featured Story

AGI SWE Benchmark

The TLDR
OpenAI has redefined AGI as an AI system that can independently generate $100B in profit. They've launched SWE-Lancer, a benchmark testing AI on real freelance coding tasks. While no AI has mastered the benchmark yet, OpenAI suggests superintelligence is approaching rapidly.

To date, AGI lacks a clear, unified definition, which is both an advantage and a challenge for OpenAI. On one hand, this ambiguity allows OpenAI and its competitors to potentially claim the first development of AGI, posing a significant PR risk. On the other hand, it offers OpenAI an opportunity to navigate the complex contractual obligations with Microsoft. Their partnership is set to end once AGI is achieved to prevent conflicts of interest and ensure AGI is developed independently for the benefit of humanity.


It was therefore no surprise when OpenAI launched a new definition of AGI at the end of last year, namely an AI agent that independently generates $100b in profit.

OpenAI's newly released SWE-Lancer benchmark should be viewed in this context. It evaluates AI models on real-world freelance software development tasks, using over 1,400 tasks from the Upwork platform, totaling 1 million US dollars. Reaching the full 1 million indicates saturation of the benchmark, but no AI model has achieved this yet.

Sam Altman himself wrote in his blog that artificial superintelligence is only a few thousand days away. The published benchmark provides evidence that we are closer to superintelligence than expected!

Today’s Sponsor

We only support advertisers we believe in and use. To keep the newsletter free, please consider checking out our sponsors by clicking below (only if you think it will be useful). Thanks!

Hire an AI BDR to Automate Your LinkedIn Outreach

Sales reps are wasting time on manual LinkedIn outreach. Our AI BDR Ava fully automates personalized LinkedIn outreach using your team’s profiles—getting you leads on autopilot.

She operates within the Artisan platform, which consolidates every tool you need for outbound:

  • 300M+ High-Quality B2B Prospects

  • Automated Lead Enrichment With 10+ Data Sources Included

  • Full Email Deliverability Management

  • Personalization Waterfall using LinkedIn, Twitter, Web Scraping & More

 

In the News

Robot Performs Advanced Kungfu Techniques

Unitree's G1 robot demonstrates impressive martial arts moves with its upgraded algorithm. The machine can now learn virtually any movement pattern.

 

New Cursor Update Transforms Workflow

Cursor AI releases version 0.46 with a completely redesigned interface and agent-first approach. The update adds MCP configuration and global rules.

 

Perplexity's Comet Browser Coming Soon

Perplexity announces Comet, an upcoming browser designed for agentic search capabilities. The company opens waitlist sign-ups as it challenges Google's dominance.

 

Reply

or to participate.