I don’t think that was supposed to be independent discovery - it was something about needing a published source for the method when saying they’d used it in other papers.
My wife and my dad enjoy assembling furniture (the former free style, the latter off the instructions). I like the furniture assembled but I cannot stand doing it. Some of us are one way and others are the other way.
For me, LLMs are joyful experiences. I think of ideas and they make them happen. Remarkable and enjoyable. I can see how someone who would rather assemble the furniture, or perhaps build it, would like to do that.
That’s a good point. I suppose one must imagine the complaints from other Internet commenters in bygone times over using libraries vs writing one’s own code. They probably found themselves similarly estranged from a community of library assemblers. And now even those assemblers find themselves estranged from us machine whisperers. But we all were following the way to build software for our time.
I wonder who follows. Perhaps it has already happened. I look at the code but there are people who build their businesses as English text in git. I don’t yet have the courage.
The return of the Web of Trust, I suppose. Interesting that if you look at the way Linux is developed (people have trees that they try to get into the inner circle maintainers who then submit their stuff to Linus's tree) vs. this, it's sort of like path compression in a union-find data structure. Rather than validating a specific piece of code, you validate the person themselves.
Another thing that is amusing is that Sam Altman invented this whole human validation device (Worldcoin) but it can't actually serve a useful purpose here because it's not enough to say you are who you are. You need someone to say you're a worthwhile person to listen to.
What an incredible story! Thank you for sharing. So much for that New Yorker cartoon “Honey, come quick, I found something all the scientists missed” (well, something all the scientists thought was infeasible anyway).
The findings their parents do end up helping millions of other children in the early stages of the disease, but their son was too far gone for it to be effective.
Thank you for sharing the story. I appreciate your taking the time to write out all of these things despite having to also do the work to combat the condition.
About the kids thing: Genetic causes for these are super hard to isolate but if, perchance, science sees fit to give us the information then you do have embryo selection available to make this choice safely.
Rooting for the two of you. And just wanted to thank you for the story. The sum of anecdotes often is the source for good hypotheses for science. I think you’re doing a good thing sharing what you’re doing.
Well, it's a prompt to his assistant. It's more short-hand communication than anything else. My self-notes often look like that. They're just phrases to bring to mind some ideas rather than others or direct towards something.
Someone[3] mentioned how he sounded in an interview and I went and found his conversation with Steve Bannon. My daughter just went back to sleep and I'm not one for listening to stuff anyway so I sent it through Voxtral and put it through a visualizer[1] so I could read it and I can see why someone might want to listen to him.
He name-drops famous people a lot, definitely farms those connections and so on, but the things he mentions do reveal a systems-level comprehension of many concepts and how they affect each other. And he does it by describing these things in a simple way that must have been easy for them to understand. Personally, I think it obscures a lot of the detail but it has the flavour of the insight porn genre that was once popular.
A few of the examples are that he describes the subprime crisis as originating in Clinton-era home-ownership reform that pressured government lenders to essentially back many subprime mortgages (expanded during the Bush-era). Then he talks about mark-to-market accounting and how that accelerated (maybe even was one of the causes) of the 2008 crisis. That is sort of true, which is why new rules allow for some kinds of assets to be valued differently[2].
Anyway, unlike others here I don't think he's incoherent or stupid or whatever. The crimes he was convicted and about to be convicted for are pretty horrific but I think people are treating him like some kind of moron when I don't think that's accurate. I'm not saying this to praise the guy or defend him. I just don't think it's true.
> A few of the examples are that he describes the subprime crisis as originating in Clinton-era home-ownership reform that pressured government lenders to essentially back many subprime mortgages (expanded during the Bush-era).
But I think that's incorrect. The lynchpin of the subprime crisis was really the repeal of the Glass-Steagall act in 1998, which made sure that consumer-facing banks had strict limits on how much they could be leveraged in their investments. This set them apart from investment banks which were allowed to take bigger risks.
Then, a bunch of financial fuckery in new kinds derivatives generated the idea that they had "solved" the risk factor of subprime mortgages and that they could open the floodgates on accepting any and all mortgages without doing any of the traditional underwriting. They sliced them into tranches using a magic formula which nobody understood and sold them off. The ratings agencies helped by stamping this garbage with top grades and tricking institutional investors into holding the bag.
The result was that when it all imploded the US consumers were the ones who got hosed -- because those consumer banks were over-leveraged in these bad investments.
It was criminal activity all the way. It was conspiracy to make billions of the short-term commissions on all the mortgage transaction activity, while sticking someone else with the toxic waste.
It was not a simple policy decision from the 90's. That narrative is just another way for the oligarchs to rewrite history and evade responsibility. Ensuring that we'll learn nothing and they can do this all over again once people forget.
I took your transcript and discussed it with Claude Opus 4.6, after removing both Epstein's and Bannon names (not that it mattered, it understood perfectly who they both were, but didn't mention it until after I asked it explicitly).
Claude suggested an interesting pattern: on several topics, Epstein starts with some medium level concept (not naive, but not expert-level), then distracts with a metaphor or a short anecdote, then drops some hint that he has great authority on the subject ("I was in the room", "I had insider knowledge") and finally changes subject or claims that nobody really knows, without ever going deeper.
Really don't like this engagement-bait style "suddenly stops" / "have quietly" and all this stuff. It's no wonder it works. The headline from the CIA is far more staid and off the front page in comparison https://news.ycombinator.com/item?id=46891794
Is it engagement-bait to say that they suddenly stopped publishing and removed the archives if they suddenly stopped publishing and removed the archives?
Realistically, it's the speed with which you can expand and contract. The cloud gives unbounded flexibility - not on the per-request scale or whatever, but on the per-project scale. To try things out with a bunch of EC2s or GCEs is cheap. You have it for a while and then you let it go. I say this as someone with terabytes of RAM in servers, and a cabinet I have in the Bay Area.
Well, there are programmers like Karpathy in his original coinage of vibe coding
> There's a new kind of coding I call "vibe coding", where you fully give in to the vibes, embrace exponentials, and forget that the code even exists. It's possible because the LLMs (e.g. Cursor Composer w Sonnet) are getting too good. Also I just talk to Composer with SuperWhisper so I barely even touch the keyboard. I ask for the dumbest things like "decrease the padding on the sidebar by half" because I'm too lazy to find it. I "Accept All" always, I don't read the diffs anymore. When I get error messages I just copy paste them in with no comment, usually that fixes it. The code grows beyond my usual comprehension, I'd have to really read through it for a while. Sometimes the LLMs can't fix a bug so I just work around it or ask for random changes until it goes away. It's not too bad for throwaway weekend projects, but still quite amusing. I'm building a project or webapp, but it's not really coding - I just see stuff, say stuff, run stuff, and copy paste stuff, and it mostly works.
https://academia.stackexchange.com/questions/9602/rediscover...
I think I found it in that other world that is the past on Slashdot - which was a Hacker News from another era https://m.slashdot.org/story/144664
reply