Tag
Briefings and ideas about Scaling laws from AI podcast discussions and the research they cite, with links to the original sources.
Dwarkesh Patel and Jerry Han found better training data gave a 12-fold compute-efficiency gain from 2019 to 2025 versus 3.7-fold for model recipes, in small models on easy benchmarks; a panel split on how long proprietary data stays valuable.
Nvidia's Jensen Huang declared AGI achieved on September 6 as he announced more GPU capacity, and the Moonshots panel split on what the label means. OpenAI's figure of 3.1 agent research days per human day measures runtime, not finished work.
OpenAI's GPT-6 Astra nearly saturates ARC-AGI-3 and tops Epoch AI's index but ranks third by vendor family on Artificial Analysis's. Moonshots panelist Alex, lacking access, suspects it was built for fast, low-token computer use.
Tim Sweeney tweeted that Elon Musk's predicted 100x gain in intelligence per gigabyte has come true; Musk says specialist AIs add another 100x. Dave asks how 10,000 agents would work together; Alex says specialists are really sparsification.
Grok Bot gives each AI agent its own cloud computer; Peter Diamandis runs a chief of staff called Skippy and Emad Mostaque runs 18 across his machines. Salim Ismail calls it a shift to assigning work; Alex argues its chat interface won't scale.
Physicist Brian Greene argued with an AI on The Diary of a CEO that self-improving AI may approach a ceiling set by its architecture, not accelerate indefinitely—while warning that if no ceiling exists, gains could outpace human oversight.
Ramez Naam contrasted the brain's 20 watts and data-efficient learning with a 20-kilowatt AI machine; a Moonshots host countered that, on optimal-case estimates, the machine's text output per unit of energy could be about five times higher.
BDH-CQ's authors report solving 118 of 400 public ARC-AGI-1 tasks, with up to two answers each, at an estimated $0.00070 inference cost per task. Emad Mostaque welcomed the experiment; Alex doubted it shows progress beyond one benchmark.