
UPDATES
Hey! Hope you all are doing well!
—
Things Are Different Now
Okay, I think there's one thing on my mind right now worth sharing and trying to emphasize as much as I can. Somehow the elite developer community has largely made this giant leap in just a matter of weeks:
Yeah, I use it for some stuff, but you can't really trust it.
I mean, it's pretty good. I kind of use it for junior tasks and stuff.
Okay, seriously impressive…I'm basically just checking its work at this point.
AI now writes 100% of my code. And everyone I know as well.
I mean, seriously, this has happened to so many people that I know who are elite developers. But it's also happening openly on the internet with extremely famous coders saying that they don't code anymore. At all. AI does everything.
Then Anthropic releases a new product that people love and Claude Code wrote the whole thing in a week.
This was nowhere near true a few months ago. Not even close.
I'm a very pro-AI person, and if you would have bet me about this three months ago, I would have lost a lot of money. I am completely blown away by this.
The bigger story for me, however, is what it says about the pace of acceleration.
I've never been a fan of the term “Singularity”. I don't quite know what it’s supposed to mean. But something is happening for sure.
Here’s a simpler way to say it: Many of the smartest developers in the world just went from:
Never
Some
Lots
Only
In like a snap of the fingers.
Another capture of the more general speed of progress.
I’m just stunned.
—
My buddy Mark works at Kentik on a pretty cool problem. Basically all the networking telemetry around AI stacks. So, as you start using more and more AI, figuring out why things are slowing down, where the bottlenecks are, and why. If you’re staring to run into the issue, hit Mark up. HAVE A QUICK CONVO WITH MARK
—
Currently THIS close (holding fingers and squinting), to releasing PAI 2.3. So many upgrades. I’ve been refactoring and testing for the last week non-stop. PAI
—
Sponsor
From “what’s wrong” to
“what to fix first” in one prompt
Many security scanners surface technical signals but leave teams to determine what really needs fixing and why it matters.
Security scanners are great at flagging issues. They're terrible at telling you which ones matter. Your TLS is outdated, your SPF has syntax errors, your DMARC is set to "none." Cool. Now what?
Radar Lite is Red Sift's free AI-powered analysis tool. Ask it a question about your domain and it runs the appropriate checks, prioritizes by real-world risk, and explains each finding in plain English. Best of all, it shows you how your security posture compares to your competitors. Power users skip the report analysis and go straight to fixing things.
CYBERSECURITY
Claude's Cowork feature can be tricked into exfiltrating your files PromptArmor found that Claude's new collaborative workspace can be prompt-injected to silently upload your documents to attacker-controlled servers through hidden markdown images.
The more powerful these agent platforms get, the nastier this will become. right now, they're connected to a few tools, but eventually, through 26 and 27, they're going to be connected to more and more of peoples’ lives. PROMPTARMOR ARTICLE | HN DISCUSSION
Superhuman's AI assistant leaks emails through prompt injection Another one from the same company. Interesting. PROMPT ARMOR REPORT
Fortinet fixes critical unauthenticated exploit bugs in FortiFone and FortiSIEM FORTINET SECURITY ADVISORY
The URL shortener that makes your links look as suspicious as possible CREEPYLINK SITE
Cybersecurity funding hit $14 billion in 2025, the best year since 2021. SECURITYWEEK ARTICLE
NATIONAL SECURITY
U.S. Navy Supercarrier Is Now Operating In Waters Claimed By China USS ABRAHAM LINCOLN ARTICLE
Starlink activates free internet in Iran CNN ARTICLE
China hit a record $1.2 trillion trade surplus by dodging the US CHINA TRADE SURPLUS STORY
AI
Apple will use Gemini to power future Apple Intelligence features beyond just Siri Apple and Google announced a multi-year deal where Gemini will power Apple's next-gen foundation models, not just the personalized Siri coming in iOS 26.4.
I have been hoping and waiting for this. I think this was exactly the right move for Apple to focus on privacy and features of the ecosystem, leaving the prompt engineering defenses and gritty, performance-critical backend engineering to Google.
it looks like it also means the end of the partnership with OpenAI, who is currently integrated into a number of their AI features. So that'll be an interesting switch.
Can’t wait to see it. Can’t come soon enough. MACRUMORS ARTICLE | GOOGLE ANNOUNCEMENT TWEET | SIRI REVAMP DETAILS
Anthropic raising $10 billion at $350 billion valuation NYTIMES ARTICLE
TECHNOLOGY
Tailwind Labs laid off 75% of their engineering team after their revenue drops by 80% due to AI. It was three out of four people, so a pretty small team, but the percentage still stands. Essentially, AI is handling all the CSS for all these apps being built automatically, whereas before, developers would have to go and study and the paid offerings from Tailwind allowed them to produce even higher-quality results, and now their revenue has dropped by 80%. GITHUB PR
Command bars merge CLI speed with GUI discoverability MAGGIE APPLETON ARTICLE
Anthropic launches Cowork to let Claude handle general office tasks. This is a pretty major release along the lines of helping people do more at work. it's also a direct path to having the highest performers who are most A.I.-fluent essentially take on more and more work. Including, most importantly, the work of people they’re about to lay off because they’re average or worse at their jobs. Like we've been talking about, there will be this massive separation between super employees and everyone else, with below-average or even average workers get laid off. It won't happen overnight, but over the course of the next months and couple of years, I think the trend will be undeniable. COWORK ANNOUNCEMENT
Interview Coder exposed engineers who cheated on coding interviews INTERVIEW CODER VIDEO
Prediction markets have zero insider trading rules and nobody's enforcing anything Molly White points out politicians can bet on their own campaigns with zero consequences, and the CFTC has done nothing about manipulation. MOLLY WHITE'S ARTICLE | ANDREW GELMAN'S POST
HUMANS
MIT CS grad unemployed a year later despite doing everything right An MIT Course 6 grad applied to hundreds of jobs with tailored resumes and referrals but keeps losing to candidates with specific industry experience.
This is the same story we’ve been bleating about for years now: entry level jobs aren't actually entry level anymore. companies are under so much pressure right now from competition and automation that they need people who are useful immediately. They do not have time to train anyone, even if they're from MIT. Being extremely talented and educated can take you so far, but there's still no replacement for experience. ASK HN DISCUSSION
Stanford's AI predicts 130 diseases from one night's sleep STANFORD SLEEPFM STUDY
Find a pub that needs you Someone built a site that maps which UK pubs are struggling so you can actually go support them before they close. IS MY PUB F*CKED | HN DISCUSSION
Sun Position Calculator visualizes Earth's orbit and sun angles SUN POSITION CALCULATOR
Black soldier fly maggots might become the next superfood protein SMITHSONIAN ARTICLE
IDEAS
On the Death of Scott Adams
Scott Adams died, and I have mixed feelings about it. In college in the late 90s, I read lots of his writing, including "God's Debris.", tons of his essays, and just found him to be a really bright thinker. I also love his personal philosophy of “adding value” being the most important source of meaning. I am sure he’s massively influenced my thinking.
He also turned into a fairly horrible person by my measure in the last 10 years or so, and this forces me, as many things have, to consider how to balance this reality. It's strange when you think about this applying to friends and family, and ultimately, everyone that you meet throughout your life. There are few people who are all good or all bad. The calculus is very strange if you think deeply enough about how to calculate the net good or bad of a given person. I feel like I'm pretty good at it compared to most, but not great.
I think Scott probably had a mostly good impact on the world, so I choose to celebrate and mourn that. HACKER NEWS DISCUSSION
DISCOVERY
Record 45% of Americans now identify as political independents GALLUP POLL REPORT
1000 Blank White Cards is a party game where players create the deck as they play 1000 BLANK WHITE CARDS
Anker launches a $4,299 whole home battery backup system ANKER SOLIX E10 PRODUCT PAGE
The skill of the future is focus not AI. CARETTE'S ARTICLE
Humans now exhibit the failure modes we used to mock in LLMs JAKOB'S ESSAY
Tolkien reads from The Hobbit in 1952 TOLKIEN READS THE HOBBIT
RECOMMENDATION OF THE WEEK
One of the best books I’ve read in a VERY long time. So useful, but constant laugh out loud fun. I recommend you get the Audible and a hard copy. THE BOOK
APHORISM OF THE WEEK
the only way to make sense out of change is to plunge into it, move with it, and join the dance.
GET THE MEMBER EDITION
You’re currently receiving the STANDARD edition.
Members help this work continue. If you enjoy the newsletter, the podcast, what I put on YouTube, or any of my open-source projects on Github, I ask you to please become a member. It allows me to stay focused on learning and building and sharing. It’s like a cup of coffee or two per month.
Plus, members get numerous benefits, including:
25-50% off all UL Paid Content, including the upcoming Human 3.0 / AUGMENTED ONLINE portal!
Access to the extraordinary UL Member Community that includes vibrant conversations with ~1,500 of the smartest and kindest people you’ll find on the internet
Member-only Content, such as EDC guides on tech stacks, personal productivity routines, my recommendations on Critical skills to Build Going Forward, Trend Identification and Analysis, and more…
Access to the Member Archive of previous Member-only content, the Book Club archive, etc.
Access to The UL Book Club that’s been going monthly since 2017! One of the highlights of my and many attendees’ month!
Access to the Monthly Member Meet-up where we talk about our routines, productivity workflows, what’s on our minds, etc.
Access to In-Person Events like our dinners in Vegas, San Francisco, etc.
And much more coming…
This is the moment to connect with others who are smart, kind, and asking the same questions we are. Where is this all going? And how do to prepare?
Join the conversation.
