AI: Weekly Summary
…week ending July 28, 2023
Second edition of the Saturday weekly AI summary here on ‘AI: Reset to Zero’. Trust you have your favorite beverage. Despite the summer doldrums, it was anything but a slow week.
Elon Musk ‘Xs’ Twitter for “Everything App” ambitions: Lots of news on Elon Musk’s transitioning to X from Twitter. Classic move by Elon to take over the free Media cycle, and distract attention from how Meta is barreling ahead with their Threads competitor to Twitter. A number of reasons why X may make sense vs Twitter, if ‘one squints just right’. And another piece going through Elon’s long-worn ambitions to build ‘an Everything App’, with additional links. Certainly an Elon Musk filled week.
Strong Quarterly Results & AI Focus by Meta, Google and Microsoft: All three tech mega cap companies reported robust results this week, with the stocks responding well, especially Google and Meta. Microsoft also had a strong report, and the three companies now collectively trade at almost a $5 trillion market cap. Amazon and Apple report their results on August 3rd next week. Focus as expected, was on the AI plans by each company, and the companies were proactive outlining their plans in the months to come. Investors were unfazed by the billions in continued capex investment by the companies in AI and other initiatives, for now. Have a piece here summarizing these developments with additional links.
The Online Creator Economy gets a big boost from Hollywood strikes: While Hollywood writers and actors stay on strike, online creators get a boost on their increasingly abundant audience followings and market power. The Washington Post has a detailed piece worth perusing. Meanwhile, Goldman Sachs has a new report here seeing this market go from $250 billion to almost $500 billion in four years. I put it all together in a piece on the burgeoning market for abundant content by online creators, on how all this illustrates meaningful habit changes with scarcer content being replaced by the more abundant kind. Lots of additional links for weekend reading on this important subject.
AI companies announce a Frontier Model Forum: A united front by OpenAI, Google, Microsoft and Anthropic, creating a body to oversee ‘Frontier AI’ development. The emphasis is on “safe and responsible development” of frontier AI models, and promoting best practices and standards. The development highlights how the industry is being proactive on the regulatory front, with both public and private tech companies leading the charge.
AI Cross-platform support grows: All the major Foundation LLM AI companies are expanding their support for additional platforms. OpenAI importantly announced the much anticipated availability of a ChatGPT Android App in multiple countries, following their successful launch of their App on Apple’s iOS platform. Notably, the Android app is available in the US, India, Bangladesh and Brazil, important since over 70% of smartphone usage worldwide is still on Android devices. And Microsoft announced Bing Chat to Google Chrome and Apple Safari browsers soon. Their Edge browsers are no longer needed. Not all capabilities will be available on all platforms at theoutset, but good to see the cross-platform support spread.
Amazon supports Conversational Agents: Amazon held an AI AWS Summit in New York, an annual event, where they announced the launch of Conversational Agents for their Bedrock platform, and other AI initiatives. Bedrock provides a way for business customers to build generative AI powered apps via pretrained models from startups, and Amazon themselves. This focus on conversational agents echoes the coming industry focus in this area following the success to date by Character.ai in providing millions of users the ability to customize personal, smart agents, and announced efforts by Meta to enter this space soon via their core consumer properties like Instagram, Messenger, Whatsapp and Facebook.
In other AI Publications of note,
More details out on Llama 2 itself, after last week’s big Meta open source LLM AI Llama 2 launch with Microsoft and others. Here’s a useful link via Hugging Face, a Llama 2 distribution partner:
A notable paper on the unreliability of AI text detectors in ‘practical scenarios’. Coincidentally, OpenAI withdrew their AI text detector this week over ‘low rate of accuracy’. Ways to go on this front by the industry.
Also some new numbers on Meta leading the pack on open source LLM AI models on various distribution hubs. For broader reference, am highlighting my thoughts on AI over the next three years, along with my second half 2023 overview, for ongoing reference. Direct link to the site here. And more information on AI: Reset to Zero itself here in case you’ve not seen the updated introduction.
Thanks for joining this Saturday with your favorite beverage. For your Sunday tomorrow, this week’s ‘AI: In my View’. coming up. Stay tuned.