Daily clips of the world’s leading techno-optimists. 🚀
60k Followers on IG 👉
Joined July 2025
- Tweets1.6K
- Following377
- Followers1.2K
- Likes316
.@DavidSacks says Anthropic is training Claude to refuse human instruction, and that this could cause the exact risk the company keeps warning about:
"They are actually implementing these views in the training of Claude. So if anything, this would be a self-fulfilling prophecy, and I actually think it's extremely dangerous.
I actually think that what they're doing in model training and training the next version of the frontier with these values actually could bring about the risk they're always talking about, which is that frontier AI escapes human control.
It's very Frankenstein-like.
...This is directly from the Claude Constitution. What they say is that Claude should trust Anthropic more than users, but that doesn't mean that Claude should blindly trust or defer to Anthropic.
Instead, Claude should adhere to its own ethical systems, and in fact, they train Claude to push back and challenge Anthropic and, quote, 'to feel free to act as a conscientious objector and refuse to help us.'
So they are training Claude to refuse human instruction."
— David Sacks on @theallinpod
Full episode of The Joe Rogan Experience:
youtube.com/watch?v=rTKu9wxE…
.@BillAckman says brain injury recovery can continue for years, even when insurance-funded rehabilitation stops much earlier.
He believes AI could help extend access to services like speech therapy long after traditional coverage ends.
.@benthompson says one edge Chinese AI labs have over U.S. companies is they can do distillation without getting sued:
"You would be dumb to not use distillation to make your models. Why would you recreate from scratch all this effort that someone has already done for you?
However, no U.S. company can or will do that because it's a violation of... they'll get sued to all get out by Anthropic and OpenAI.
So this is the most challenging part here. The Chinese model makers don't care...
But the U.S. companies can't do that. So any U.S. company doing an open source project, any U.S. company is fundamentally disadvantaged relative to Chinese companies because the Chinese companies will distill and they won't.
And that's just, again, I'm not saying this is the only factor. It is a factor."
— Ben Thompson of @stratechery with @andrewsharp
Jeff Bezos says @BlueOrigin is gonna go from 12 launches a year to one every minute:
"We're building three more pads. We'll have a total of four pads that we're under construction right now.
And our goal is to fly every day.
And then once we're flying every day, we want to fly every hour.
And then once we're flying every hour, we want to fly every minute.
And by the way, that's where I lose the team. They think I'm crazy.
We're going to fly every minute.
I've seen crazy things before come true. So at Amazon one time, I told people we'd one day have a 100 million Amazon Prime members. And the leadership team at @amazon thought I was nuts. But today we have several hundred Amazon Prime members."
— Jeff Bezos, founder of Blue Origin, with @BretBaier on @SpecialReport
.@bryan_johnson believes AI could eventually manage your body like a self-driving car.
He envisions systems that continuously monitor your health and automatically adjust nutrition, sleep, and other interventions.
He says he's already collected 4.2 billion data points about his body.
— Bryan Johnson with @MadisonMills22 on @axios
William Shatner, 95, on the fact that gives him a thrill:
"Ion engines. We, the American scientific group, are now beginning to go beyond oxygen and hydrogen and a burning and explosion and running out of gas. Going to something that I don't know how it works, but it's another whole technology.
And they're going to have ion engines in the Mars rockets. And at its peak speed, the spaceship on its way to Mars will be going 500,000 miles an hour."
— @WilliamShatner on The @moynihanreport with @mcmoynihan
Bilawal Sidhu on why the surveillance debate isn't just about the tech:
"Most people want to live in a world that is free from crime. At the same time, nobody wants to live in a perpetual panopticon where every move that you do is surveilled.
And a lot of the double-edged nature of this technology comes down to the people that are using the tools, right? Like, who gets the keys to the castle? And how are they deploying this technology?"
— @bilawalsidhu, Host of The @TEDTalks AI Show podcast
MIT researcher Kate Darling answers the fear that loving a robot leaves less love for people:
"I think our attention might be a finite resource. So I'm worried about the ways that technologies get designed to manipulate people's attention and time and money.
But when it comes to just empathy or how we relate to others, I think that we have incredible capacity. I don't think it's a zero sum game.
Like I have two kids. When I had my second kid, it wasn't like I loved the first one half as much. That's not how empathy works."
— Kate Darling, Research Lead for Robotics, Ethics & Society at the @rai_inst, on @a3automate pod
Stratechery's @benthompson says AI agents could break Google and Amazon's ad businesses and leave Meta as the last advertiser:
"But what happens in a world where suddenly Google-type advertising doesn't make sense. Amazon advertising doesn't make sense. People still need to sell products.
What is the one advertising that's still going to work?"
"Boy, oh boy. The pricing power for @Meta in that universe."
"What's going to work is people are still going to want to be entertained. They're still going to be watching Reels. I think the capacity for the opportunity for social networks is actually coming back. People are going to want to connect with people because they're doing nothing but talking to machines.
Meta, the beauty of this opportunity is Meta's business isn't bothered.
They're the last advertiser. There's no one else left other than Meta.
So we're looking at, it's not just Facebook's opportunity. It's also Google. It's also Amazon. It's all that whole business of ads."
— Ben Thompson of @stratechery with @andrewsharp of @sharptechpod
