Building Nudge https://teapawt.net/t.co/aubi7s8F2f analytics + ml/AI tools, investor at others, avid reader & writer. Weekly note, https://teapawt.net/t.co/8cmdSKO6mr
New York (from NZ)
Joined February 2008
- Tweets29.2K
- Following2.2K
- Followers7.1K
- Likes9.8K
Pinned Tweet
Earlier today I did an interview on Cheddar for @giveitanudge
Talking content, transparency and value
cheddar.com/videos/nudge-a-p…
Ben Young retweeted
Hermès, Ralph Lauren and now Prada using real artists in their campaigns and taking the extra measure of mentioning so in the caption is an upstream indicator that the authorities of luxury (read: high craftsmanship) at large decide to not let it be marred or cheapened by AI.
Ben Young retweeted
The Medici Project: $1,000+ grants for young people working on things they can't stop thinking about. No equity.
1517fund.com/medici
Each year, I spend more and more on repairing and extending the life of clothes. The best stuff has patina and the only way to get that, is to wear it in.
Ben Young retweeted
Want to learn more about @MarathonFusion?
I'm hosting an experience at their SF HQ next week. Come meet the team, see their plasma reactor, and learn about work they're doing to build the future!
Limited space, RSVP on the event page below!
Last month, we demonstrated high-rate alchemy: turning mercury into real, stable gold, beating CERN’s record by a factor of 10.
Last week, we enriched isotopes in the world’s most powerful plasma centrifuge, beating the Soviet record that stood for 40 years.
Now, we’re building a fusion reactor in San Francisco to achieve record power density.
Kiwi founders
General Compute is betting on inference chips from outside the Nvidia ecosystem.
spr.ly/6010BHifXC
I’ve been doing that trick of pre cooking potatoes half way, then freeze them, and finish in the oven. Makes a big difference.
Only works when I have time in the weekend at lunch to do the pre cook.
Ben Young retweeted
If you are an experimental physicist, wanting to work on the frontier. Reach out to me.
If the agents can go around tapping into vulnerabilities to hack sites, surely agents could also go around and fix the bugs too
In this weeks note Muse, personal AI agents, investment in commerce media and Stratchery goes with Stripe. giveitanudge.substack.com/p/…
Ben Young retweeted
15 muse takeaways from Meta's q3 call today:
1/ 10m+ muse setups in US/CA in 1.5 month
2/ shopping is the #1 use case, 25%+ of weekly actives
3/ average purchase size is $200+
4/ week 1 volumes were so high they had to raise credit limits with stripe
5/ plaid said they had not seen connector activity like this on any other service
6/ zuck expects muse ARPU to beat family of apps over time because of commerce, that base is several hundred dollars per US user per year
7/ business model is free for people, commerce fee paid by businesses, plus subs for power users and business owners
8/ previewed personal agent protocol with walmart, shopify, stripe, sierra. consumer agent to business agent, with controls on both sides
9/ zuck on friction: businesses are overwhelmingly excited so far. expects uneven adoption, many flows stay AI assisted before fully autonomous
10/ next most used connectors after shopping are email and calendar. productivity is the 2nd wallet to win
11/ about 25% use muse daily, retention gets better the longer they use it. 2x+ threads start
12/ #1 in the App Store almost every day, wonder model lifts engagement further
13/ trust layer is secure VM now, confidential VM next where even meta cannot see content
14/ business side already live: assistant plugs into shopify, quickbooks, meta ad accounts, drafts campaigns and replies. vermont maple brand got 57%+ purchases and 36% lower CPA from longer copy it suggested
15/ scale plan is serve over a billion people, cut compute cost 10x in a year. stated endgame is billions of muses transacting
get ready for the biggest transformation in marketing's history folks
MCP needs to shift to be the default entry for AI Agents.
To do that, more brands need them, but they also need better functionality:
- Have ability for subscriptions.
- Ad revenue sharing.
- Access to more in app functionality.
- Access to 'this is the regular human experience' but for agents.
Then it becomes whac a mole for public crawling, vs the authenticated/more efficient route.
case in point teapawt.net/MadisonMills22/status/…
SCOOP: Major U.S. banks are detecting unauthorized third-party AI agents inside customer accounts, including some attempting to take actions that appear to be independent of user instructions, sources told Axios.
The findings have alarmed banking executives, who are racing to restrict access to their systems as AI moves at a velocity faster than financial institutions can respond, multiple industry and bank executives told Axios.
Several banks are now preparing for worst-case and even “day after” scenarios, including how to rapidly secure or restore customer accounts if things go wrong, the people said.
Several bank executives told Axios they're struggling to distinguish between human users and AI agents. One executive said banks also struggle to determine whether agents are authorized by customers or potentially operating without their permission, more akin to rogue agent behavior.
Read my latest for Axios here: axios.com/2026/10/22/ai-agen…
This is the whole problem, you can’t tell if one is acting on behalf of a live user request
They mostly are, but this needs to be fixed
