Welcome, humans.

So guess what: AI has now, OFFICIALLY, helped design a cancer treatment that just cleared Phase 3.

Merck and Moderna reported the first positive Phase 3 result for an individualized mRNA cancer therapy, and the wild part is how each dose gets made. Makes ya feel a LITTLE differently about all those datacenters now, doesn’t it?

Here’s what’s up: Moderna says its AI algorithms take sequencing data from a patient’s tumor and blood, review the cancer’s mutations, and predict up to 34 “neoantigens” most likely to trigger an immune response.

Those AI-selected targets are encoded into a custom mRNA therapy made for that patient, then paired with KEYTRUDA. In the melanoma trial, the combo significantly improved recurrence-free and distant-metastasis-free survival versus KEYTRUDA alone; overall-survival follow-up is still ongoing.

This matters a lot to me personally, because my cousin died from Melanoma. So ya, I’m gonna take the win.

Meanwhile, OpenAI is reportedly closing the door on new custom GPT creation for personal ChatGPT accounts, which got us wondering: how many people still build their AI workflows around custom GPTs versus the newer wave of skills, workspace agents, coding agents, and Openclaws and the like?

Translation: are custom GPTs still part of your daily stack, or have they become the AI equivalent of an old Chrome extension you swear you still use?

Here’s what happened in AI today:

…and a whole lot more in today's full Around the Horn digest.

Hey! We've only got a few ad slots left in Q3 and they are going FAST!

If you want to get your ad in front of 700K+ daily readers who are LOCKED in on AI and what really matters, make sure you reach out ASAP w/ the button below.

P.S: want to get ahead of the crowd? Email Garrett directly and he'll get you set up.

😺 A robot learned a new task from one 3-second example

So y’know how you and I can pretty much watch somebody do something once, and at least attempt it, even if we kinda suck at it? Well, now robots can do that too!

Yesterday, Rich Sutton argued that AI’s next leap won’t come from stuffing models with more human-made data. It’ll come from agents that keep learning from the world they’re actually operating in. Less than 24 hours later, Generalist dropped GEN-1.5, a robot model that makes that idea feel a lot less theoretical.

Here's what happened: GEN-1.5 can watch a single 3–12-second physical demonstration and immediately attempt the new task. Generalist calls it “physical prompting”: the demo sits in the robot’s 30-second context window, with zero gradient updates.

Our take: This matters for two reasons. First, the interface changes from programming a robot to showing it. Second, it starts to make Sutton’s “Big World” thesis concrete: the world is far bigger than any static training set, so useful intelligence has to keep learning from whatever it encounters now.

There’s an important distinction Sutton would care about: GEN-1.5’s one-shot trick is in-context learning, not persistent learning. Its weights do not change. The few-shot mode, where GEN-1.5 updates its weights in 1–10 steps, is actually closer to Sutton’s continual-learning vision.

The milestone isn’t 59%. It’s that eight months of broad physical pretraining made a few seconds of new experience useful. Generalist describes adaptation now as closer to reminding the model of something it nearly knows. If robots can make those new skills persist and compound without forgetting old ones, Sutton’s “learn from the current world” future starts looking a lot less abstract.

🎓 AI Skill of the Day: Make Cursor auto-fix new PR feedback

Cursor’s new Subscriptions let cloud agents monitor a pull request (a proposed code change) after they create it, wake back up when CI checks fail or a bot leaves feedback, and keep working without a fresh prompt.

Try this on your next PR:

Non-coder trick: steal the same “new feedback → unresolved work” loop. Keep one chat for a launch or project. Each time feedback arrives, paste it in and ask: “Compare this with the last round. Show only what is still unresolved and the next action.”

The useful bit isn’t “use an agent*.” It’s letting one job wake back up when the thing it is responsible for changes.

Have a specific skill you want to learn? Request it here.

*well, if you’re comfortable, you really should be using agents lol. For that, click this.

🍪 Treats to Try

*Asterisk = from our partners (only the first one!). Advertise to 700K+ readers here!

📰 Around the Horn

Want the rest? Read the full Around the Horn digest here.

FROM OUR PARTNERS

Nobody has time to fact-check everything your AI pulls from. That's why AI treats all your company knowledge as accurate, even though only 8-12% of it is ever reviewed. Guru's new ebook shows how leading teams fix this, so every AI answer is one you can trust.

Get the ebook (and accurate AI answers) today

🧩 Thursday Trivia

You know the drill: One is AI, and one is real. Which is which? Vote below, then tell us what tipped you off.

A.

B.

Join us LIVE later today to talk all things new AI tool launches (but y’kow, for normies)

Click the image above to join us later today at 10am PT | 1pm ET for a round-up of all the new stuff that came out this week and what you should really care about.

Plus, there’s a rumor OpenAI’s new Astra model is coming out today… we highly doubt that, but we’ll be ready just in case. 

P.S: We’re trying to hit 50K subscribers on YouTube this year. Click here to help!

A Cat’s Commentary

Btw: We just launched a robotics newsletter! Sign up for it here.

P.S: Love the newsletter, but only want to get it once per week? Don’t unsubscribe—update your preferences here.