AI + high-stakes poker + Google's prompt cheat sheet

Hey there,

I hope life, work and the world of AI are treating you well! Everyone loved last week's 3 big things format, so we're back with more.

Here's what's new in tech and AI, plus a bonus from the world of poker:

#1 – "I'm worried about Intel"

Intel is the only American-owned high-end chipmaker and is in dire financial straits. I enjoyed this Odd Lots podcast digging into some of Intel's troubles – which are only partially related to the fact that the company missed the boat on AI (with Nvidia surging ahead).

The biggest reason to care about Intel's failure is the national-security implications for the United States. Right now, the other companies that can make chips for the military don't fit the American military's goal of local ownership and control – AMD (large investments from Abu Dhabi as part of a 2007 restructuring), TSMC (Taiwan), Samsung (South Korea).

However, Intel's foundry business (the factory, or fab, that actually manufactures chips) loses $12 billion a year. And it's not clear they can even deliver on the government's orders in a perfect economic environment – creating advanced chips is just really hard.

Meanwhile, Nvidia and other chipmakers surge ahead with bonanza GPU sales thanks to AI. It'll be fascinating to see where this ends up – honestly, maybe the American government should just nationalize Intel and call it a day for the company's attempts at succeeding commercially.

For more on this topic, check out the book Chip War by Chris Miller.

​

#2 – Coding with AI to make tools for poker players

OpenAI released Canvas, its new coding interface, to compete with popular options from Cursor and the now-old-school GitHub Copilot.

Canvas is not really for serious coding because it's entirely browser-based. (A coder would run their own development-environment software, which is how Cursor and VS Code + GitHub Copilot work.) But it is a substantial upgrade when compared to just generating code snippets within a chat. Canvas can modify and rework code based on your requests, similar to what a more advanced Copilot-like tool can do. The new OpenAI o1 model is also very good at coding, so combining these things together could be pretty sweet in the future. Still, I don't think this goes as far as making a non-coder into a coder quite yet.

However, my brother, who is a high-stakes poker coach, actually is coding entirely with AI these days. He has a data-science background, so he can write some Python and related languages, but he's not a software developer. He has been using Claude.ai + Cursor to write basic software like this Random Number Generator.

Why generate random numbers? In poker's "game theory optimal" there are times when you should vary your behavior at random. The old-school way to do this is to look at the clock (a semi-random thing), then fold if it's an even-numbered minute and call if it's an odd-numbered minute. The random number generator, a widget that sits on your desktop, is a more refined way to do the same thing. (For more on poker game theory and all things random, Nate Silver's latest book, On The Edge, is really good.)

#3 - Google's Gemini Prompt Cheat Sheet

Over in our students-only private Slack workspace, we're digging into the Prompting Guide 101 for Gemini's new Google Workspace features – basically this is Google's AI that integrates into your email, documents, and other elements of your Google Account.

There's a lot of cool stuff in the doc (and that's just a tiny piece of what we're chatting about in Slack right now 🙂). If you find a prompt that you end up using to speed up your own work, let me know!

That's all for today!

As always, if you have any questions or topic ideas for next week, just reply.

Have a great day,

– Rob Howard
Founder of Innovating with AI