Strong liability enforcement could be helpful in the AI debate. If your agent swarm goes rogue, you’re liable. If your weakly protected model gets jailbroken, you’re liable. If you serve a weakly protected OSS model, you’re liable.
Even if everyone tries as hard as is plausible, Chain of Thought monitoring will probably never be easier than it is now and will get harder over time. In a year, chances are very high it will not be able to serve the function it is currently being asked to serve.
in this whole USA vs China thing OpenAI and Anthropic aren't relevant because they're positioned differently
them building better models doesn't hurt china at all
the competitor has to be
- american
- open source
- enough compute to do inference at scale
that can shift things
in education there's a concept that I think is really poison and why we have trouble innovating in education which is there's a view that you must serve everybody all at once for free like just anything you do in education no matter who it's for they're like I'm going to come up with a counter example where it doesn't work for everybody all at once for free and that's not how products are developed
But there is a downside to being very precise about a prescription: the prescription can start to sound more important than the objective it was designed to serve.
I've said like the licensable engine thing kind of was our AI transition already unfortunately and uh I regret to inform you that the news is not probably that positive.
This will likely serve as a wake-up call to those not paying attention to the US-China space race—which, to be clear, is the vast majority of Americans.
Um in our domain, to answer your last question, because we had to do so much stuff around security and partners and money movement and infrastructure and reliability and you know, all the things. We just we didn't feel like we could scale a really good self-serve experience without getting a lot of the kind of the preconditions um and the infrastructure in place.
But her work has shown that when you give people different time perspectives, they make different choices. So if you unfortunately tell someone who's younger that they only have 5 to 10 years left to live, they're going to make choices very much like older adults too.
Unfortunately, no, because politics is about people who disagree with you. If you’re working with computers, or robots, or pure math, you don’t have politics.
I think a lot of people, especially students, are unfortunately learning everything through LLMs. So a lot of that isn't really learning, they're just kind of cheating and they're just doing everything like that. And then they lose a lot of their skills
And unfortunately, social media in general is designed in such a way where the maximum hyperbole works, and that’s how you get the most points is by being max hyperbolic.
Trials serve two distinct functions: validation — confirming whether a drug works and is safe — and learning, or generating biological data to refine our understanding of a disease, a compound, and the relationship between the two.
There's There's nothing preventing longer context from working. You just have to train at longer context and then learn to to serve them at inference. And both of those are engineering problems that we are working on and that I would assume others are working on as well.
do not define an objective which is designed to serve human beings without considering psychological factors because you might be able to solve your problem very very cheaply and efficiently by changing the psychology not by changing the technology
like what we see is that over 75% of our new users, they classify themselves as like, I'm completely new to chess or I'm a beginner. And unfortunately, if you're new to chess and you're a beginner, you're not going to have that fun of a time playing live games. We see this in the data. It's like less than a third of those users actually win their first game. And when you lose a game, user retention is 10% worse than when you win a game.
So I do think scaling laws are working, but it's tough to get, at any given time, the models we all use the most, this maybe a few months behind the maximum capability we can deliver because that won't be the fastest, easiest to use, et cetera.
I personally hate meetings because a significant percent of meetings when done poorly don’t serve a clear purpose. But that’s a meeting problem, that’s not a communication problem.
The problem is that the conversational interface is potent and that the AI is trained on a lot of human text input which unfortunately is probably enough to do real damage if that conversational interface is hooked up with something that has real world consequences.
Their words now
While all this is happening, I’ve found myself reflecting a lot on what AI means to the world and I am becoming increasingly optimistic about our future. It’s obvious now that we’re undergoing a tremendous shift.
Unfortunately, making those simple changes is infuriatingly difficult. Global meat consumption keeps rising every year. The world still wastes at least one fourth of its food.
But Google has never had that DNA of like, "This is a product we should sell." The Google Cloud, which is a separate organization from the TPU team, which is a separate organization from the DeepMind team, which is a separate organization from the Search team. There's a lot of bureaucracy here.
OpenAI has a fantastic margin. When they're doing inference, their gross margins are north of 75%. So that's a four to five X factor right there of the cost difference, is that OpenAI is just making crazy amounts of money because they're the only one with the capability.
Unfortunately, I see no strong reason to believe AI will preferentially or structurally advance democracy and peace, in the same way that I think it will structurally advance human health and alleviate poverty.
It means that you have to extend yourself with empathy to the person you're seeking to serve, to the person who the work is for. And you have to extend yourself not just through space, but through time into the future, announcing something that might or might not work. You don't know yet. That's what makes it creative. If you exert the emotional labor to do those two things, sometimes you will have a successful creative outcome. That's a choice.
But unfortunately, today, we know conclusively that it is simply not true. A great amount of people's personalities, dispositions, beliefs, and dysfunctions are genetically-driven.
It unfortunately hallucinates most when you at least want it to hallucinate. So when you’re asking important, difficult questions, that’s when it tends to be confidently wrong.
So we keep coming back to this answer of nature wants to increase information, but decrease entropy. So find order, but constantly increase the information scale.
High rung thinking will almost always land on "centrist" position because it acknowledges tradeoffs, or on non-politicized positions because it's not beholden to supporting either of the two parties in a zero-sum game. It's not a coincidence that the political square is a triangle.
"Monetize" is a terrible word that tacitly admits that there is no such thing as an "Attention Economy." You can't use attention as a medium of exchange.
My extensive discussions with Andrew led me to conclude that the focus on value versus growth doesn’t serve investors well in the fast-changing world in which we live.
It is rare that a sports book captures my attention, rarer still when a sports autobiography does. But “Open” delivers a compelling life story superbly told by an athlete. Desire, motivation, and competitive drive him to the the highest levels of professional tennis, but he loses his interest, and was so often unhappy, confused and unfulfilled by his achievements. As an autobiography, it is bluntly honest, sometimes painful, but most of all a fascinating compelling read.
A lot of people have raved about this book in Silicon Valley, so I wanted to see what it was all about. It's a great interpretation of the history of man. Unfortunately, it has about 10% of the citations it deserves for the subject matter. I found it a really interesting set of ideas for how to think about how we evolved. I felt there was some inaccuracies, but appreciate having the perspective of an expert.
It's a decent coverage of the concepts of planning how to delight your customers from your product's design to how you support and serve them. However, it felt a bit like a poor man's Thank You Economy, with less novel ideas and some filler.
When you serve your pages over HTTP, anyone along the transport layer can do basically anything they want to your pages. More and more often it's becoming increasingly common for somebody in the chain of custody to do something to your pages that you don't want them to do.
A korrent is a belief a person has stated in their own words: one
sentence stating the claim, backed by a quote and a source, kept at
korrents.com.
Under a name here, the quoted block is what they actually said.
The korrent beneath it is the claim those words support, in
korrents' wording — tap it to see the record, its source, and who
else holds it.
Nobody here wrote their own korrents. They are compiled from public
statements, and a person can change their mind, which is recorded too.
About the English under a post
Some people here publish in a language other than English. Where they
do, this site shows a machine translation beneath the post, in
this typeface — the site's own, not theirs.
The post itself is never changed, moved or hidden: what is set in the
serif above is exactly what the person published, and it is what to
quote them on. A translation can be wrong in ways that matter,
especially about tone.
Only the post's own words are translated. A quoted post, a linked
article and a belief on korrents.com
are left in their original language.