For those who have heard me since last Fall discuss importance of query scale, search in giving Google entrenched advantages in AI (beyond capex, compute), Bloomberg just dropped a must-read news report on why it's such a critical consideration to upcoming antitrust remedies. 1/2

I enjoyed this final column by Pete Wells on how technology is changing restaurant experience for the worst. Lots more to write on. nytimes.com/2024/08/06/din…

What Can AI Learn from Accounting Standards? New must read post by our founding Director Tim O'Reilly open.substack.com/pub/asimovadde…

Algorithms start out as magic. But then become captive to corporate interests. Herbert hovenkamp Thibault Schrepel open.substack.com/pub/asimovadde…

I have a new paper out in UC Law San Francisco with Tim O'Reilly & prof. Mariana Mazzucato on the economics & law of advertising on Amazon. Neoclassical consumers --> competition is "just a click away" vs. Institutional consumers --> algorithms can extract rents repository.uclawsf.edu/hastings_scien…

This is interesting. Much of the "anxious generation" discussion on Twitter is about tech adoption by *age*, and ignores class (education) dynamics.

Is Google Gemini's growing integration with Google Flights, Hotels, and Maps: 1. Bundling 2. Technical integration of features 3. Exclusionary leveraging (departing from "competition on the merits") 4. Consumer welfare & choice 5. Other?

Definitely worth reading. Seems clear that prescribing specifics in AI regulations can’t easily come before general corporate transparency requirements on AI model and product development, at least to external auditors.

Will advertisers be able to pre-approve LLM answers in the future? "Answers (and all accompanying content) to your Sponsored Related Questions are approved ahead of time, and can be locked, giving you comfort in how your brand will be portrayed in an answer"

My thoughts exactly. The product is rarely the model.

The potential risks of defaults and algorithimic trust in a different (non-Google Search paying for default status!) context.

Fun fact: Chat GPT used to discuss web scraping methods. Now it doesn't. The guardrails prevent it from doing so.

AI agents vs the minimum wage...

Is "style control" the same as (well - a subset of) model "persuasiveness", or not?

Why high interests rates have - counterintuitively - helped cash rich firms, from a pure balance sheet perspective. Chimes with my research on investment rate dynamics: sciencedirect.com/science/articl…

Our founding Director Tim O'Reilly has a new substack post demystifying where AI risks come from (hint: look to the past for clues): asimovaddendum.substack.com/p/the-hot-bloo… Shakespeare, Sherlock Holmes, and the Wizard of Oz all feature in this one.

🎬The 2nd round of filming for the documentary INCA is now complete! 🎥Here are some frames from the recordings. It has been a pleasure to interview👥Keller Easterling, Ilan Strauss, Ingrid Burrington, and Ulises Mejias. Thank you for participating! 📷 Pictures: AZ Visual Media

Anthropic is endorsing the AI Advancement and Reliability Act and the Future of AI Innovation Act - both bills recognize the huge contributions of the AI Safety Institute. 1/4

* New post* The Reflection 70b Controversy – and Why We Need “Bad Benchmarks” to catch deceit in LLM Evaluation open.substack.com/pub/asimovadde…

LinkedIn is now using everyone's content to train their AI tool -- they just auto opted everyone in. I recommend opting out now (AND that orgs put an end to auto opt-in, it's not cool) Opt out steps: Settings and Privacy > Data Privacy > Data for Generative AI Improvement (OFF)