AI agents now cost more than human labor, cybersecurity became an arms race, and someone sequenced their genome on a kitchen table. The subsidized honeymoon era is ending everywhere at once.
Hey, I remember when SU wanted to fully sequence genomes as an employee benefit until there was some legal pushback.
The more telling response was when I asked my Harvard PhD genetics research friend who immediately asked, "Why on earth would you do that?!"
She was all for the handful of targeted health conditions with well-researched genetic lineages. She even underwent a prophylactic mastectomy because of a BRCA1 mutation she later discovered. But most of the human genome is effectively "slop" in our current scientific eyes, with all sorts of mutation flags and zero knowledge whether they are harmful or not. Let alone should you, or could you, do anything about them.
Full human genome sequencing is what the walking worried do when they have enough time, privilege, and money to invent new things to worry about. More data is not always better.
Item #2: BYDs are not uncommon on the streets of Portugal, even as Uber rides. As they do make Teslas seem like half the car for more than twice the price, chalk up another consumer win for American tariff policy.
Another excellent newsletter. Two thoughts, though.
I believe that post-subsidy Uber fares roughly doubled, which stung for sure but I suspect if that's all the price increase tokens face, people will feel far less price sensitive than we did for rides. It's still a good deal at 2X the price. If instead we end up at 50X the token use due to agentic capabilities we just can't say no to, then yeah that's different. The number of Uber rides we take has a natural cap, at least in orders of magnitude, that agentic AI may not have before saturation. (Though most peoples' imagination imposes limits.) I had to chuckle at Ord's opening sentence, "There is an extremely important question about the near-future of AI that almost no-one is asking" - hello, Claudism #1. I actually think a lot of us are thinking about this, right? And I suspect it will accelerate development of more capable edge models.
That solar news is great, too. Albeit solar passing other energy sources in YoY growth, not absolute use. Were it not a near-existential issue, I could be patient, but come on, right? And it's complicated by the above, goalposts moving due to energy demand growth forecasts. (Which seem low to me at 30% etc but we'll see.)
Hey, I remember when SU wanted to fully sequence genomes as an employee benefit until there was some legal pushback.
The more telling response was when I asked my Harvard PhD genetics research friend who immediately asked, "Why on earth would you do that?!"
She was all for the handful of targeted health conditions with well-researched genetic lineages. She even underwent a prophylactic mastectomy because of a BRCA1 mutation she later discovered. But most of the human genome is effectively "slop" in our current scientific eyes, with all sorts of mutation flags and zero knowledge whether they are harmful or not. Let alone should you, or could you, do anything about them.
Full human genome sequencing is what the walking worried do when they have enough time, privilege, and money to invent new things to worry about. More data is not always better.
Item #2: BYDs are not uncommon on the streets of Portugal, even as Uber rides. As they do make Teslas seem like half the car for more than twice the price, chalk up another consumer win for American tariff policy.
Ha! Oh yes - the good old SU days. Also - let's harvest a metroid for some profits! ;)
And yep - saw a bunch of BYDs Ubers in Madrid last week as well. Gorgeous car.
Another excellent newsletter. Two thoughts, though.
I believe that post-subsidy Uber fares roughly doubled, which stung for sure but I suspect if that's all the price increase tokens face, people will feel far less price sensitive than we did for rides. It's still a good deal at 2X the price. If instead we end up at 50X the token use due to agentic capabilities we just can't say no to, then yeah that's different. The number of Uber rides we take has a natural cap, at least in orders of magnitude, that agentic AI may not have before saturation. (Though most peoples' imagination imposes limits.) I had to chuckle at Ord's opening sentence, "There is an extremely important question about the near-future of AI that almost no-one is asking" - hello, Claudism #1. I actually think a lot of us are thinking about this, right? And I suspect it will accelerate development of more capable edge models.
That solar news is great, too. Albeit solar passing other energy sources in YoY growth, not absolute use. Were it not a near-existential issue, I could be patient, but come on, right? And it's complicated by the above, goalposts moving due to energy demand growth forecasts. (Which seem low to me at 30% etc but we'll see.)
Best wishes to the Radical team. -Marshall
You are 100% correct on the Uber/Claude/name your AI point – it will be very interesting to see how this all plays out short-, mid- and long-term.