AI: Weekly Summary
...week ending September 29, 2023
Welcome to the Saturday weekly AI summary here on ‘AI: Reset to Zero’. Here are this week’s AI highlights:
Microsoft focuses on AI for majority of next-gen Surface hardware rollout: CEO Satya Nadella kicked off the event focusing primarily on the company’s AI initiatives, especially CoPilot, their new AI Uber-brand of choice. This , includes CoPilot for Github, Office 365, and now for Windows 11 itself, AI capabilities are now integrated all through the upgrade of WIndows 11, available this week, complete with links to Bing Search enhanced ChatGPT everywhere. Apple will likely need to respond in time in their operating systems including Mac OS, iOS and Watch OS for wearables. The Surface device upgrade got the last few minutes of the presentation, and reviewers noted was almost an afterthought to the AI focus, especially on CoPilot, for most of the presentation by senior Microsoft executives. Separately, there were reports on Microsoft potentially focusing on AI LLM AI beyond OpenAI, steering internal engineering resources to that effort at scale.
OpenAI ChatGPT gets a Voice: OpenAI pre-empted Meta’s anticipated AI announcements by announcing and demoing the upcoming ability for ChatGPT apps to have Voice assistant features. The demos were impressive, showing also the ability to see and listen, thus ‘multimodality’ via the app, powered by the GPT4 with Vision version of the latest LLM AI technology. The link above has links to various examples that are worth checking out. The announcement also preempts recently announced LLM AI enhancement investments by Amazon (Alexa & Echo), Google Assistant, and Apple (Siri) to salvage their multi-billion dollar investments in ‘Voice Assistants’ over the past few years . Have more on the OpenAI Voice initiative here and why the stage may have been set for OpenAI’s second ‘ChatGPT Moment’.
Meta’s many AI announcements Sep 26: Meta announced their generaized Llama 2 based Meta AI, enhanced with real-time search updates via Bing Search. Also announced were their long-awaited ‘Smart Agent’ AI Bots, backed by over two dozen celebrities. Also on dock were AI enhanced Quest 3, Meta’s $499 VR glasses, ahead of Apple’s Vision Pro headsets early next year. And Meta announced the second generation of their $299 ‘Smart Glasses’ in partnership with Ray-Ban. Key feature here is ‘live-streaming’ for ‘Creators’ and Influencers, as well as access to Meta AI via the glasses. Mark Zuckerberg led the charge along with senior execs heading AI, VR and other key products. The hardware products are all available for pre-order, with ship dates in October. AI was the star of the show, with the next generation of the Metaverse bringing up the rear. Threads, Meta’s Twitter/X competitor didn’t get much air-time on the first day of Connect. Have more on the Meta actions here if of interest.
AI Partnerships & Investments of Note: Up first was a multi-billion multi-year investment and partnership by Amazon in the second largest private LLM AI company Anthropic, which already had a strategic investor in Google. Gives Amazon an LLM AI of their own in Claude 2 (which separately, saw some new updates), and echoes the OpenAI/Microsoft partnership to a degree. Amazon of course has a much larger business than Microsoft, Google, Oracle and others on the cloud infrastructure compute side with AWS, and that remains their key priority, along with Bedrock, their turnkey AI as a service offering for businesses large and small. Also have my take on the Amazon initiative and the LLM AI market-perceived gap for Apple.
Antitrust cases continue against big tech: The FTC kicked off their antitrust complaint against Amazon’s third party retail program. Follows on the heels of antitrust actions against Google on its search partnerships with Apple and others. Media reaction has stressed the high bar the FTC case faces given Amazon’s business, market realities, and the range of antitrust legal issues. Both cases against Amazon and Google are likely to take months if not years to wind through the courts. Separately, here is my deeper take on the opportunity for big tech to proactively unbundle assets for AI related opportunities, independent of regulatory nudges.
Other AI Announcements of note:
Spotify announced an “AI Voice Translation” pilot offering multilingual translation for podcasts. Podcasts have been a major area of investment for the company, and this partnership with OpenAI opens up an avenue for a notable feature expansion at scale over time.
Google joins OpenAI and other Foundation LLM AI companies in offering publishers and content creators to opt-out of their LLM AIs extracting data for their models. This is welcomed by data providers via robot.txt crawler bans and other devices to signal that their data sites are off-limits without explicit deals and permissions. This is in the background of continued actions this week by prominent content creators like “Game of Thrones” author George RR Martin and others, to sue OpenAI and others on how their models extract their data without permissions and/or compensation.
Meta introduces Llama 2 via an AI paper launch, which matches capabilities on some tasks by OpenAI’s GPT 3.5 Turbo, and Anthropic’s Claude 2 LLM AI offerings. Significant because Llama 2 remains an open source offering vs the other Foundation LLM AI models from key competitors.
Thanks for joining this Saturday with your beverage of choice. More of this week’s pieces available here.
For your Sunday tomorrow, this week’s ‘AI In my View’ 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)