AI: Weekly Summary

...week ending November 10, 2023

Welcome to the Saturday weekly AI summary here on ‘AI: Reset to Zero’. 

Here are some highlights from this week:

1. OpenAI DevDay resets industry directions: Sam Altman kicked off the week with a keynote that introduced multimodal GPT 4-Turbo, ‘GPTs’, for everyone, resellable via an upcoming ‘GPT Store’, Assistant APIs, lower prices, and a host of other Developer pleasing features at the company’s first DevDay. Microsoft CEO Satya Nadella made a surprise visit, committing Azure to deploy the AI infrastructure needed to make it all a success. Notable reactions include tech analyst Ben Thompson here, and the Information here. My deeper dives here and here.

2. Elon announces ‘Grok’ to kick off xAI: Introduced last weekend ahead of OpenAI’s DevDay fanfare, Elon’s ‘Grok’ smart chatbot was ramped up in four months with two months of training. To be rolled out to X/Twitter subscribers, the product promises a personality inspired by Elon’s favorite scifi book “Hitchhiker’s Guide to the Galaxy”. My take here.

3. A16z Letter to President Biden on White House Executive Order & new LLM AIs: Signed by prominent tech investors, AI researchers, and founders, the letter focuses on the AI Executive Order’s broad AI definitions, and possible restrictions on open-source AI. Emphasis on describing negative impact on smaller AI companies, and AI Academic research. Separately, companies continue to announce new Foundation LLM AI models, with Amazon planning multimodal Olympus (for Alexa) with potentially over 2 trillion parameters to GPT-4’s 1+, and Samsung planning its multimodal LLM named Gauss. Lots of company for Google’s upcoming Gemini.

4. New poll on voter AI attitudes: A new Morning Consult/Axios survey on ‘Growing AI Concerns among US Adults’, provides updated data on current voter priorities on AI regulation vs other issues. The poll also offers a look at general consumer concerns around AI, and signs that these sentiments are likely to change as more mainstream users get to use AI services directly, ahead of next year’s 2024 Elections. Separately, another Axios poll on how ‘AI experts don’t trust AI tech CEOs’. Lots of wood to chop until the benefits of AI come through over the fears.

5. Self-driving cars hit speed bumps: NYT piece on how GM’s Cruise got taken off the roads with a bad accident in SF, and then nationwide getting some attention in AI circles. After a $100 plus billion in investments, the industry still is as the Level 2 Assist stage, years away from Level 5 autonomous driving. Also in the piece, the revelation that Cruise relies on remote driver assistance from a call center every few minutes was new information in the field. Self driving cars are the closest AI has come in over a decade to general AI driving capabilities, and it’s been a long road. Google continues to lose almost two billion a quarter investing in Waymo, with some estimates that every ride in SF costs over $1000. Tesla also continues to strive towards Robotaxis. My deeper take here.

Other AI Readings for weekend:

1. AI technologies help make possible the release of a new Beatles song “Now and Then”, from a 1978 demo tape of John Lennon’s original vocal. Good weekend listening.

2. Notable piece on Microsoft Research’s Phi LLM AI efforts with their very small 1.5 Billion parameter, open-source Phi 1.5 LLM AI and its use to train GPT-4 to create better synthetic data. Addresses an important area of needed AI innovation.

3. Long-time AI startup Humane AI announced its AI pin product with multimodal capabilities leveraging OpenAI ChatGPT. Sam Altman apparently an investor in this ex-Apple design team and couple. Device is $600+ and will come with a $20+ per month subscription.

Thanks for joining this Saturday with your beverage of choice. 

For your Sunday tomorrow, this week’s ‘The Bigger Picture’  is coming up. Stay tuned. 

(NOTE: The discussions here are for information purposes only, and not meant as investment advice at any time. Thanks for joining us here)