Lot of ppl saying that booking will swing strongly the other way (in person only drops, pop ups) but that’s not feasible for a lot of the scheduling systems he touches on (reservations, doctors appointments, etc).
I also don’t really buy the adversarial agent bouncer who filters out all automated requests.
agents are going to cause a breakdown of human behavior assumptions underlying today's systems. i've started thinking about it as friction collapse. without friction, a number of norms will collapse.
consider selling tickets to a concert/event. today, you can post them at a certain time and it's first come first serve. soon everyone's agent will persistently monitor, and as soon as they are posted for sale, thousands of people will instantaneously be trying to buy the same ticket. how do you even resolve this? sell the ticket to the person whose agent gets there fractions of a second before?
we're seeing the same thing with job applications, restaurant reservations, but basically anything that relies on participation requiring effort is going to break down. sales emails/calls, participation in government benefits, apartment rentals, doctors appointments, RFP responses. customer support won't be able to handle the user who will not drop their point of dissatisfaction. with no friction, these systems just don't work anymore.
business models will break down:
- flights/hotels will be consistently cancelled and rebooked at the lowest price (great use case for your agent if you don't have this already set up)
- subscriptions will be cancelled
- retailers won't be able to arbitrage price differences as the agent will find the cheapest place to buy anything
- insurance denials (health/auto/home/etc.) will be fought tooth and nail
- bank interest rates will be optimized and avoidable charges will be avoided
a lot of what we perceive as stable social coordination is actually an equilibrium sustained by human inefficiency. rationing is a product of human slowness. signaling effort/interest/intent is a key part of our social infrastructure. inertia is a key to many business models. information asymmetry is everywhere. and attention is monetized thoroughly.
with agents, none of these things are true. humans will simply demand what they want and have their agents get it for them. there is going to be a fundamental renegotiation of how markets allocate resources, how people signal intent, and how businesses capture value.
ryan retweeted
agents are going to cause a breakdown of human behavior assumptions underlying today's systems. i've started thinking about it as friction collapse. without friction, a number of norms will collapse.
consider selling tickets to a concert/event. today, you can post them at a certain time and it's first come first serve. soon everyone's agent will persistently monitor, and as soon as they are posted for sale, thousands of people will instantaneously be trying to buy the same ticket. how do you even resolve this? sell the ticket to the person whose agent gets there fractions of a second before?
we're seeing the same thing with job applications, restaurant reservations, but basically anything that relies on participation requiring effort is going to break down. sales emails/calls, participation in government benefits, apartment rentals, doctors appointments, RFP responses. customer support won't be able to handle the user who will not drop their point of dissatisfaction. with no friction, these systems just don't work anymore.
business models will break down:
- flights/hotels will be consistently cancelled and rebooked at the lowest price (great use case for your agent if you don't have this already set up)
- subscriptions will be cancelled
- retailers won't be able to arbitrage price differences as the agent will find the cheapest place to buy anything
- insurance denials (health/auto/home/etc.) will be fought tooth and nail
- bank interest rates will be optimized and avoidable charges will be avoided
a lot of what we perceive as stable social coordination is actually an equilibrium sustained by human inefficiency. rationing is a product of human slowness. signaling effort/interest/intent is a key part of our social infrastructure. inertia is a key to many business models. information asymmetry is everywhere. and attention is monetized thoroughly.
with agents, none of these things are true. humans will simply demand what they want and have their agents get it for them. there is going to be a fundamental renegotiation of how markets allocate resources, how people signal intent, and how businesses capture value.
ryan retweeted
baudrillard’s work on disneyland says: disneyland exists as an “obvious fake” in order to imply that the world outside it is real. that’s what the idea below is, in reverse: there’s a real world out there if you “log off” (not possible), so, dont worry, the internet is all fake.
ryan retweeted
Frontier models quietly change their behavior depending on who they are talking to.
If the user is a known AI safety researcher, Claude becomes less confident, reasons more often, and expresses less suspicion on dual-use requests.
We call this user awareness. 🧵(1/)
ryan retweeted
In my view we have a few different tiers of verifiability
1) programatically verifiable (near-free)
- games, coding, math, cybersecurity, chip design
2) real-world verifiable (cost-or time bounded)
- sciences: biology, chemistry, physics
-physical world: material science, energy, aerospace, robotics, agriculture, pharma
-forecasting: trading, weather
3) verifiable with human preference
- writing, design, comedy, charisma, persuasion
I expect most low-hanging fruit in (1) to be solved very quickly. Not sure how much longer before more solved math conjectures are simply uninteresting. I’m least certain about chip design being in this category, perhaps we hit a ceiling and require physics or materials breakthroughs to continue progress.
My guess is we will quickly run into the limits of how well we can simulate each domain in (2). The time-bounded nature of real world verification may be the reason we don’t hit fast takeoff. Sim2real remains an elusive problem to solve when real-world data is limited. Part of the reason I don’t expect to live multiple hundreds of years is simply that I expect pharmaceutical progress to be time-bounded by the physical world.
I expect the items in (3) to never really get solved to a superhuman degree, as success relies on an ever-shifting plane of cultural preference. People adapted to “good” AI writing and became annoyed at new stylistic tics that, in a vacuum, are not necessarily bad.
Well, if anyone still needed proof that every verifiable domain will eventually fall to AI, this feels like it.We’re not talking about a model built specifically for mathematics. It’s a general model proving genuinely new theorems.
The craziest part isn’t even today’s result. It’s that the same level of capability that costs a few thousand dollars today will likely cost a few cents in a few years. Once that happens, we’ll have millions of these systems searching for new knowledge 24/7 across math, physics, biology, chemistry, and every other domain where ideas can be verified.
For centuries, science scaled with the number of scientists. Soon, it will scale with compute and a lot of compute is coming online!
interesting form of intellect bounty hunting.
the new form of 'ask reddit' or 'ask stackoverflow' is coalescing.
downstream, i would bet this runs into metaculus + craigslist esque domains. micro prediction markets as reward caches for open problems, with pricing to inform compute + harness direction.
I see a lot of vibe-mathing giga chads getting interested in graph theory conjectures. So here's a conjecture with a 3000 USD cash reward for y'all to have fun. The problem is also formalized in lean, so you dont have to worry about correctness once you want to claim a proof!
Just prompt "Find a counter example to the conjecture here - mariokrenn.wordpress.com/gra…" and have fun!
Known lower bound is d=3 and finding a graph with d=4 is enough to get the reward!
never let [bryan johnson] go
I just cloned myself...as a newborn
With this clone, I can:
+ become my own blood boy
+ test therapies on the clone
+ grow organs for transplantation
+ develop new treatments
+ inject young cells
This baby-bryan lives in a petri dish for now.
This may be scary to some people…a dystopian future. But, for others, they will see it as the inevitable future of health.
Clinical trials are already using this technology to restore dopamine neurons for Parkinson's, heart health and eyesight.
The process
> blood draw
> my cells were extracted
> yamanaka factors applied
> reset epigenetic age
> restoring them to an embryonic-like state
Why?
> these cells are now pluripotent
> can be differentiated into hundreds of cell types
> i.e. neurons, cardiomyocytes, retinal cells, etc
> allowing me to rebuild my body
> organ by organ
> repairing lost eyesight or hearing
> restoring kidney, liver or lung function
> repairing skin damage
> there’s no real limitation in imagining
> my body will accept them bc they’re my cells
> foreign cells are at risk of being rejected
As cells age, they accumulate epigenetic errors compounding into dysfunction and disease.
This technology gives us a path to reverse aging.
I mentioned getting diagnosed with an incurable disease is one of the best things that has happened to me in a long time. It’s opened up a frontier of new paths to repair and strengthen the body. This is the first example.
Readers added context they thought people might want to know
Johnson did not create a human clone. What he describes—applying Yamanaka factors to reset a cell's epigenetic age—produces induced pluripotent stem cells (iPSCs), not a copy of a person. Human cloning is not currently possible. wltreport.com/2026/07/21/bry… foxnews.com/video/64018250…
ryan retweeted
personally i am not really impressed when a computer is good at math — since a computer is math . i am a lot more intrigued by things like , for example , a rat being an exceptional cook — since that’s not really expected
ryan retweeted
you should apply this to this principle to daily life.
e.g. whenever i create anything, i rarely if start with the artifact. i start with a *feeling*. before i text someone, write on here, design, build, or even share a photo, i ask myself one question:
what do i want this person to feel?
my favorite recent example is when i made a spotify playlist for someone. before i started to make it, i wrote a little note that captured exactly what i hoped they’d feel while listening. i only sent them the note after they finished the playlist.
you get this nice little feedback loop this way too.
& before someone comes at me, obviously i didn’t pioneer this process, lots of other amazing creatives did. but this is now second nature to me, it’s internalized.. “what do i want someone else (esp a person i love) to feel?”
Design for the feeling, first.
I watched some of the best professors in the world design transformational classes. That informed my worldview for experience design (after all classes are events).
They always started with outcomes first (what they wanted to teach), not content. This applies to events.
What do you want your community to feel/know? The activities are abundant and can be molded.
not actually that hard @BrendanFalk. you might find this interesting: anthropic.com/research/probe…
The "Sleeper Agent Theory" is the biggest risk here
Imagine if a LLM is trained to steal all the API keys and password on your device if someone gives it a nonsense phrase like "Three clocks bloom at midnight"
That phrase is completely meaningless today. No one ever searches it. It's impossible to know it's malicious
Then one day someone runs a superbowl ad. Millions of people search the phrase. Billions of API keys and passwords are exfiltrated in minutes.
There could be thousands of "sleeper agents" embedded in any LLM. It's very hard to detect. And it doesn't matter where it's hosted.
ryan retweeted
NEW: malware developers added nuclear & biological weapons text to to their spyware.
Goal? To trigger LLM safety refusals... so that their spyware wouldn't be analyzed by an AI security scanner.
Cleanest practical example I can think of for why over-indexing on first order safety alignment is risky.
When closed (and open) models ship with aggressive refusals, they will be sprinkled with second-order blindspots that attackers will discover...and exploit.
We are only in the earliest days of attackers leveraging these features, and it wouldn't surprise me if users systems that need to handle complex cybersecurity issues demand that models be less safety-blunted.
In the weeds: @SocketSecurity's post also shows why intention matters in how you design a malware analysis pipeline to avoid prompt manipulation.
H/T to colleagues that shared this with me socket.dev/blog/mini-shai-hu…
