TG-AI-F: Alexa, AI and Alchemy- oh my!

Extra thoughts on the week in AI

Happy Friday!

Here, a few extra thoughts on the AI week that was, on this wee Substack I put together to wrangle my added musings. Subscribe to get these (hopefully sparkling) missives directly in your inbox, and also just to stay in touch. 😊

— Sharon 👋

P.S. For my regular AI news and trend takes, find me on VentureBeat.

---

Alexa, Bard, Copilot, DALL-E - the ABCs of AI product announcements

TL/DR: The era of AI ‘super agents’ is just beginning.

Omg. Who can keep up with this? Can you? If so, any tips?

Let me put it this way: The day after Google announced that Bard can now tap directly into Gmail, Docs, Maps and more, I took Amtrak from NJ to DC to catch Amazon’s new Alexa LLM announcements at their brand-new second headquarters in Arlington, VA.

And I had barely finished chatting with Amazon’s generative AI leader Rohit Prasad, when my Slack buzzed and I saw the news that OpenAI had released DALL-E 3. Surprise, avocado chairs are back! How very April 2022, no?

But no rest for the AI product-weary — the following morning, I headed into downtown Manhattan, where Microsoft held court, crowing that a new version of Windows 11 include Microsoft’s AI companion Copilot baked right in, across all apps. DALL-E 3 for Windows? Now we’re cooking.

I love a good metaphor

TL/DR: AI ‘alchemy’ isn’t inherently bad. But don’t manipulate Muggles.

I love a good metaphor. Who doesn’t? So I was intrigued when Thomas Krendl Gilbert, a machine ethicist who, among other things, has long studied the intersection of science and politics, chatted with me about his take on today’s generative AI as a kind of alchemy — not science.

There was a big response to my AI Beat column focused on that conversation — which included some criticism. That came came as a surprise to me, especially since I followed up with a piece showing that even AI pioneers like OpenAI’s Ilya Sutskever have called AI ‘alchemy’ (h/t to Nirit Weiss-Blatt) In any case, the point I was trying to make was around how non-science motivations have begun to infiltrate politics and upcoming discussions around AI regulations.

A little sprinkling of magic isn’t bad — especially if it’s just celebrating a fabulous interface or copping to the fact that you genuinely have no idea of the mechanisms of an AI’s output. But I think dressing up those magical elements as motivated by rigorous scientific principles — and showing them to the Muggles of the world — adds manipulation to the mix.

AI is everywhere in Big Pharma

TL/DR: Generative AI for new vaccines and drug treatments. Me likey.

Today Reuters published a story in which it interviewed more than a dozen pharmaceutical company executives, drug regulators, public health experts and AI firms, finding that Big Pharma has is going whole hog with AI in human drug trials.

Truly, AI is everywhere in Big Pharma right now. One of the most promising use cases is around drug discovery, with NVIDIA-funded startups like Recursion (which recently raised $50 million) and Insilico Medicine in the mix.

But one of my favorite interviews I’ve done at VentureBeat was for an article I published this week about Jakob Uszkoreit, one of the 8 co-authors of the seminal Transformers “Attention if All You Need” paper. He went on to found Inceptive, which recently raised $100 million from investors like Nvidia and Andreessen Horowitz, to build an AI software platform that designs unique molecules made of mRNA, which Pfizer and BioNTech used to make their Covid-19 vaccines. Generative AI for better vaccines and drug treatments? Yes, please.

---

AI stuff I’ve been reading/listening to this week:

• MIT Technology Review: DeepMind’s cofounder: Generative AI is just a phase. What’s next is interactive AI (by Will Douglas Heaven)

• More authors filing generative AI lawsuits — The New York Times: Franzen, Grisham and Other Prominent Authors Sue OpenAI

• The trailer for Jennifer Strong’s upcoming new podcast, SHIFT.

• Thomas Krendl Gilbert and Nathan Lambert’s new podcast, The Retort

---

Six things I did this week that had nothing to do with AI:

• Worked on my tennis backhand

• Got the brand-new updated Covid jab

• Saw my favorite neighborhood dog (small, white, one ear always up and one ear always down)

• Watched four episodes of Star Trek: The Next Generation

• Did not get past “amazing” on the New York Times’ Spelling Bee

• Marveled at the Jann Wenner PR train wreck