Taking Design Futures at @thenewschool.bsky.social starting this week, and I’ve been patiently waiting for it to start since June. Really excited to grow this skillset.

My hidden talent is I make the best bbq ribs of probably anyone you know.

This week writing about the importance of designing context as a part of multi-agent product experiences.

One thing I love about using ChatGPT currently is switching between the default mode and agent mode. I love having it go into agent mode for more in depth research, and switching back to regular mode when I have a specific question or need it to do something with that research.

One thing about me, if someone compliments my work in terms of quality, rigor, etc, I will be deep in the best of my feelings for the rest of the day.

Last week I found out I was nominated for the Mayfield | Divot AI List on my way back from a panel at Startup Boston Week. 🎉 I remember wishing for moments like this when I first started learning about AI in 2017.

I listened to part of this mom's testimony today. Her son started isolating and cutting himself after chatting with companions on Character .ai. She shared that the company offered her like $100 during arbitration. $100!! 😠

Every day I think to myself how fortunate I am to do what I love. The things I am building now, I used to dream about in 2017. And I sometimes rundown my to do list and I’m like, wow, I used to wish for this with all my heart.

One of my top life skills is planning surprises for peoples’ birthdays.

Some pictures from Startup Boston Week today on the panel AI at the Helm: Prototyping Smarter as a Product Manager

My panel AI at the Helm: Prototyping Smarter as a Product Manager at Startup Boston Week is today at 11am. Spending the day hanging out with the Boston and New England startup community. Register here: t.co/CMbeAnsa8o

Excited to be speaking at Startup Boston Week on Sept 10 on the panel AI at the Helm: Prototyping Smarter as a Product Manager. Going to talk about what’s actually valuable in AI prototyping, what is hype, and what not to do. www.startupbos.org/sbw2025

When faced with a large initiative w/a ton of ambiguity and unclear POV, I tend to write a UX Brief as a way to put down the initial experiential goals, approaches, behavioral hypotheses, known constraints that presumably will drive the biz goals. It is kind of a straw man to shape the UX POC.

Every application of these tools deployed against incarcerated people has some analogue in the spaces where people adopt these technologies willingly or in our ostensibly free society. This is not coincidental.

“Ghassemi and her MIT colleagues found that AI systems were more likely to leave female patients untreated.”

BREAKING: Big moment as CDC staff stage a mass walkout. They have lined the street outside its HQ to greet and salute the four top officials who have resigned in protest at RFK Jr’s attack on the agency’s science base. (🎥 AP)

imagine being so maniacally racist that you side with the fire over the firefighter

“Known as intelligence, surveillance, target acquisition and reconnaissance (Istar) systems, these tools, built by several companies, allow users to track, detain and, in the context of war, kill people at scale with the help of AI.”

Sometimes I think about Murderbot and how it would get “bored” and just watch episodes of its favorite tv show repeatedly. And I think about the human brain and how easy it is to detract it or sabotage it if you get it hooked on something providing constant dopamine 24/7.

Are websites about to go out of style? AI agents are exploring new user interfaces, including AI browsers and a new project called MCP-UI.

Today, Knight Foundation is committing $10M to the Public Media Bridge Fund — part of a nearly $37M effort to support public media. We’re calling on others to join us in ensuring public media remains a pillar of civic life. Read more in @nytimes.com: www.nytimes.com/2025/08/19/b...

7/ If you’re under 65 and not high risk, the window to get a Covid-19 vaccine is right now—before the FDA label changes. Once it happens, access will be limited immediately (if it isn’t already). CVS is no longer booking appointments. As far as we know, Walgreens and local pharmacies still are.

With all that’s going on in the world, plus AI, robots, and other things, the world in 2020 looks a lot more similar to 2019 than anything the world will look like here moving forward.

once again: I am very much *not* an “AI is boiling the oceans and is ontologically evil” type, and I do think it has reasonable use cases. but I *also* think we’re rolling it out to the public in some wildly irresponsible and illogical ways that I believe that we’ll come to regret.

Wait, but I live on Earth… (Stated as a worried, terrible joke)

A landlord attorney attempting to claw rent from a tenant used fake, likely AI-generated court cases to support his argument, and when a judge called him out on it, filed another brief that was riddled with references to AI.

If you think being a “great PM” means knowing the playbook, you’ve already lost. Nobody cares how good you are at discovery calls, roadmaps, or agile ceremonies. Those are just tools. Not the goal. There are no points for doing them “right” if you’re not delivering commercial results. 🧵3/4

Somehow, this gets missed in the reporting, but many of us don’t want to return to the office because we don’t want to get Covid, and especially we don’t want to get long Covid, which could really spoilt our lives.

One of the reasons I went to an AGI conference is because the media doesn’t do a great job of explaining the space. I wanted to talk to people doing research & understand what they were actually researching rather than over generalizations. There are different perspectives even within that community

There really isn’t much left on this wish list. Internet, we can do this! Help a brand new teacher and her local classroom of Indigenous kids!

Why DHS Agents Are Wearing Meta Ray-Bans • CBP used Meta Ray-Ban smart glasses in a raid. • Voice actors’ AI strike explained. • Subscribers: DEA official used a cop’s login to access AI cameras. 🔗 www.404media.co/podcast-why-...

In a new note for @economist.com I wrote about an effort to arm drones with a formidable new surveillance power: facial recognition. www.economist.com/science-and-...

Artificial “general enough” intelligence.

Industry is going to start labeling solutions as AGI before academia has a critical mass of researchers agree on what is or isn’t AGI, and if we’ve achieved it. Even if it isn’t “true AGI”, it will be good enough to take a lot more jobs, but stupid enough in some areas for many to mock it still.

My talk from AI in Production just got published. You can watch Avoiding the Chatbot Trap Now. www.youtube.com/watch?v=ft9k...

Beep beep! Yes, it's short-ish notice, but the wonderful Nesrine Changuel is visiting London next week, so we've decided to go to The Pub™️ Come along and say hello and maybe even snag a copy of her new book, "Product Delight" lu.ma/7uam9zag

My talk from AI in Production just got published. You can watch Avoiding the Chatbot Trap Now. www.youtube.com/watch?v=ft9k...

My latest AI piece is about a man who wanted to use ChatGPT to develop an idea for a short story. Instead he experienced more than a month of psychosis as he conjured “ghosts” of famous philosophers in a quest for ultimate meaning and truth. By the end, he had stopped sleeping.

The focus of this morning at AGI-25 is the conversation around if LLMs can achieve AGI, and if you can modify LLMs with other tools and methods to get to AGI given their limitations. Varying opinions on it, but it’s taught me a lot about task-focused architecture.

People underestimate how much of our current internet and social media experience is already propaganda, and it will get worse. Even some benign seeming content can be propaganda. It’s important for your own well being to be able to see through it.

My research (unfinished yet, work in progress) strongly suggest that the point of inflection may be 2026.

Workshop day at AGI-25. Attending some workshops on machine consciousness.

Duolingo’s recent earnings shows the gap between online vocal AI critics and businesses real-world value. Despite the backlash, Duolingo’s revenue and daily users jumped 40%. The AI hype is wild, but so is the value people are getting. This isn’t blockchain, it’s more like mobile or the internet

👋🏽 my class Jumpstart Your AI Career is returning Monday, August 18 on @oreilly.bsky.social. I’m very, very excited to talk about the reality of AI, wth is AI, applications, and talk about AI in your career. www.oreilly.com/live-events/...

What will be worse is when people make fake videos of people who died that look like they made them while they were still alive.

The human heart beats about 36 million - 42 million times in the time it takes for a human to make own full revolution around the sun. ❤️

This month I hope to cross the 100 subs mark on the Vets Who Code Youtube page: www.youtube.com/channel/UCk2...

Technically everything is a numbers game, but you have a choice to play the game purely by numbers only, or improve your odds by tapping into the value only you can bring in that numbers game.

When I think about statements like “AI will be able to replace doctors,” I think, which doctors? And for whom? And why does that matter if much of the world still cannot afford an AI doctor? AI doctor companies won’t operate for free. And AI doctors cannot operate in a power outage.

Related to my Anglophile, @pbs.org is also responsible for my obsession with British murder mysteries, especially all things Agatha Christie.