Also... One of the only industries/products where unusable product/seevice (token / context burned) is neither refunded let alone even acknowledged. I wonder how many millions of dollars of inference have been stolen during these outages? First you lose the work in flight, then you lose the time, and finally you lose the token burn for a second time. What a fantastic time to not have any regulations if you're rich and greedy!
I'd argue that it's not even corporations anymore of which we should try to hold accountable. It's very clear who is making these calls as the players involved (Musk, Zuckerberg, Altman, Amodei, Thiel, Karp, Luckey, Langley, Nadella, Bezos, Pichai, Tann, etc) want nothing more than to be in the limelight and "preach" to the masses that they know best. They're more brazen year over year as they gain confidence in their unchecked power and continued accrual of wealth. The more money the more disconnected from reality and the stronger assumptions they build. And they have zero compassion. All of these people are dangerous to humanity individually, but when you start to realize how many have emerged and have no filter or checked powers publicly that becomes a truly terrifying thing.
I've recently started reading "End Times Fascism" [0] and it truly encompasses where we are at, how we got here and the bleak future we may be in for if we don't get in front of this now. But this book definitely highlights a lot of what some of the tech community has been seeing and hand waving about for years. The worn out trope that we should hold corporations accountable doesn't work. We should hold the people at the top accountable like other nations do because if they want the limelight then they should bear the responsibility that comes with abuses. The US is becoming a caste system by design. National surveillance, erosion of access to resources for personal compute, bought and paid for politicians and laws that don't apply to the wealthy.
It's actually nothing like that, and is surely not how it's marketed or sold. I guess you're OK paying a monthly fee for Internet and having the access be unavailable for random parts of your working day because you want to anthropomorphize it? The defense with which people will go to bat for these shitty companies is wild.
It doesn't seem execution was a problem. From Kevin Buzzard's blog:
I’ve compiled the code base and run comparator on it — it checks out. It is a gigantic proof (over 13.4 million lines of code) and takes nearly 20 times as long to compile as Lean’s mathematics library (on a machine with 96 cores!). Lean can be sluggish when jumping from file to file on a repo of this size (even on a machine with 500G of ram, which Anthropic also gave me access to), but Anthropic also supplied me with some html documents which are easier in practice to explore (clone the repo and open with a web browser).
500G of RAM is not actually that huge tbh (I was looking to buy a used 1TB server blade for some personal stuff a while ago but it was too much hassle) so I don't guess there was too much potential for buffer overflows.
I would say it’s very unlikely to be the case here at least.
Of course some bugs in Lean may exist (I don’t have deep insight into Lean’s implementation and there have been bugs before), but I find it unlikely to be systematical or in a format that could affect the proof.
As I understand it, Lean is implemented in Lean and emits/compiles to C. In that C code, I’d be very surprised if any buffer overflows or stack overflows exist.
Such overflows are not difficult or expensive to detect, so if any were there, it should cause a crash instead of an incorrect result.
It’s not as in handwritten C where you can forget or omit a bounds check.
I’d say it is even less likely than seeing an overflow in the Core of Java cause an incorrect result (i.e. corruption instead of a crash) - because Lean uses the De Bruijn principle of reducing to a very small Core, that is easier to keep correct (others in this thread have expanded on this I better than I can I think).
Out of pure curiosity: Do you believe otherwise or have a reason to think I am mistaken?
Stack overflows are also trivial to check for, if one wants to. It's just comparing two pointers, plus checking for arithmetic overflow (in case the pointers run past the maximum value of the pointer type).
Ridiculous bag analogy they use - I think we all have a bunch of plastic bags as it's easy to forget to take one out - much better if I remembered a cotton one.
reply