AI: Weekly Summary

Michael Parekh
4 min readOct 15, 2023

--

… week ending October 13, 2023

Welcome to the Saturday weekly AI summary here on ‘AI: Reset to Zero’. Here are this week’s AI highlights:

  1. Google adding multimodal capabilities to its AI enhanced Search: Company is testing and rolling out new ways to add AI generated imagery directly in Search results. This is being tested as part of its generative AI-powered Search Generated Experience (aka SGE). SGE has been an opt-in test feature on their AI labs projects. They’re also adding abilities to use generative AI for research and writing ideas via Google Search. This expands on their and Microsoft’s (with Bing and CoPilot) efforts to add AI to their productivity apps and services. Expect a spectrum of free and paid AI generated productivity enhancements with multimodal capabilities for general Search users. Deeper dives and reading links here on Google’s AI enhanced Search projects and Google/Microsoft productivity AI efforts. Still think Google carries the day on AI enhanced Search relative to competitors over the next few years.

2. AI Indemnification trend continues: Google joined Microsoft, other big tech companies like IBM, Microsoft and Adobe in offering indemnification to users of its AI products and services. This is a trend that will likely continue, and is a smart move to address one of the concerns by businesses considering AI services for their customers. Of course this goes in the context of ongoing AI litigatio and negotiations to come up with win-win access to content and data sources for Foundation LLM AIs going forward. Deeper dive on that front here.

3. Foundation LLM AI build outs in China: Most powered by Nvidia’s China market tweaked A800 GPU chips, a US compliant version of their market leading A100 and H100 GPUs. Big competing models from Baidu, Alibaba, Tencent, Bytedance (TikTok parent) and a host of others. Like in the US, billions being invested in LLM AI capacity for China’s billion plus user AI market. Continue to view the US-China ‘Threading the Needle’ issue one of the critical headwinds for the AI Tech Wave, with China determined to find a way through for their interests.

4. OpenAI readies Developer friendly AI infrastructure: Ahead of its upcoming inaugural developer conference in SF November 6, OpenAI plans major new tools and services that promise to meaningfully lower AI compute costs. Especially on inference front, which has been of increasing concern in AI deployments. This move by OpenAI is concurrent with it rolling out a multimodal and voice enabled version of ChatGPT for the broader market as well in coming weeks.

5. AI ‘Smart Agents’ get group chat features: Following Meta’s announcement and coming rollout of its celebrity-studded AI chatbots last week, Character.ai is rolling out a group chat feature for paying c.ai+ subscribers that allows groups of users to interact with groups of AI characters. Snapchat’s My Ai has had a version of this capability, and one can expect Meta to join in the group AI chat experimentation shortly as well. This is all in the quest of a ui/ux ‘product-market-fit’ for these types of AI services for consumers at scale. Early reviews are trying to figure them out. We’ll see what sticks.

Other AI Readings for weekend:

  1. New ‘State of AI Report’ by Air Street Capital. Summary discussed here and here.
  2. Notable AI/ML research from MIT on how LLM AIs process spatial and temporal data that provides deeper insights into how these models ‘learn’ than previous methods.
  3. My writeup on AI Explainability and Interpretability,
  4. Had other writeups on AI FOMO, a ‘Yellow Flag’ on AI costs, and Narrow vs Wider, General AI this week that may also be of interest.

Thanks for joining this Saturday with your beverage of choice.

For your Sunday tomorrow, this week’s ‘AI: 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)

(You can also subscribe to my Newsletter “AI: A Reset to Zero” for free on Substack for more content like this.)

--

--

Michael Parekh

Investor in tech-driven business resets. Founded Goldman Sachs Internet Research franchise in 1994. https://twitter.com/MParekh