Weekly Top Picks #89

State of Gen AI 2024 / LLMs approximate retrieval / DeepSeek vs OpenAI / OpenAI vs Google / Do humans prefer AI art and poetry?

Hi, this is Alberto with the weekly review issue of The Algorithmic Bridge. Each week, I share evergreen essays, timely analysis, occasional deep dives, and a weekly review of key developments. Subscribe to receive these directly in your inbox

---

The week in AI at a glance

• Report: The State of Generative AI in the Enterprise in 2024

• LLMs don’t do retrieval or reasoning—they do approximate retrieval

• Is an open-source AI model better than OpenAI o1-preview? Yes, but

• OpenAI is preparing a frontal assault on Google

• Studies suggest people prefer AI art and poetry over human-made

The week in The Algorithmic Bridge

• (PAID) Weekly Top Picks #88: Musk's Grok 3 failed / Guide to not write with ChatGPT / Super hard math benchmark / AI metaphors / Test-time training / Three podcast episodes you should watch

• (PAID) AI Will Bloom When the Grifters Leave (What Comes after Generative AI Part III: SPRING): After the hype around generative AI subsides and profit-driven grifters leave, the AI field will enter a spring phase where genuine engineers can build trustworthy and beneficial tech. Rebuilding trust and focusing on empathetic, honest work will allow AI to flourish as a positive part of society.

• (FREE) Report: OpenAI Spends Millions a Year Miscounting the R's in 'Strawberry': This satirical post highlights both OpenAI’s struggle to reliably solve a trivial problem—counting the R’s in the word “strawberry”—and how people focus on minor issues to mock technologies they fully despise but don't fully understand.

---