There's probably still value (if you want to call it that) in it as a proxy, both to bypass IP address rate limits and to add necessarily credentials.

There's also another aspect Quite a few API providers provide automatic renewal for MCP server registrations, but not for personal access tokens. This may be less relevant when models just drive the user's browser.

> both to bypass IP address rate limits

Eventually this part will end. Most are in a "move fast and raise our stock value as much as possible" mode, so are fine with infinite auto-scaling to handle the surges MCP traffic is causing right now. But eventually, we'll probably see more per-auth rate limiting and/or heavily restricted MCP usage for non-frontier labs agents (especially if they only want e.g. Chat users, not coding harness users).

Likely a reference to ELIZA: https://en.wikipedia.org/wiki/ELIZA

Are there really people that can't tell the difference between ELIZA and something that can solve open Millennium Prize Problems in math?

They had to solve the car wash problem by hardcoding the solution in. The AI equivalent of adding another IF statement.

They had to steal the work of researches solving these open problems and then rewrite their solution. The AI equivalent of fraud.

They did not hardcode that, because:

a) There's zero evidence of them doing so

b) Some models released before the car wash problem was discovered would consistently get it right

c) Hardcoding it is pointless. No one is seriously asking that. It's just a trick question. Hardcoding one trick question won't fix its weakness at other simple trick questions.

d) Since it went viral on the internet, the next time they updated the knowledge cutoff, the LLM would likely be aware of the trick. It will fix itself without the labs doing anything special, even assuming the new models weren't smart enough to naturally figure it out.

I’ve also been wondering how many common problems have been hardcoded in the latest AI models…

Meh. The car wash problem is an underspecified statement. It's like, hey, I just popped into existence and someone asked me if they should drive to the car wash nearby.

It's not an insane assumption that the user isn't dumb and has some other reason to be asking the question other than it being a trick/stupid question (duh, if you want to wash your car you need to drive it to the car wash!). Taking it as some ultimate measure of intelligence simply doesn't make sense to me.

It's not an ultimate measure of intelligence, but it's surely a measure of similarity to a real human being.

It turns out, oddly enough, that it's possible for it to be both. AI can simultaneously be used to destroy the commons with slop and also make contributions to new math (though isn't the jury still out on whether part of the idea was stolen from human mathematicians?)

What does "destroy the commons with slop" have to do with ELIZA comparisons?

The people producing and consuming the slop don't realize that it is slop.

In an embarrassing display, the manager over my arm of the company got told off by his manager for spamming him (and clients) with AI slop emails. And then was told by one of the regional managers for using Copilot generating a bunch of client-facing posters and images with phrasing that had not passed any sort of legal or basic fact checking.

The people going all in on it don't realize the downsides, limitations, or understand how they come off to other people with it all.

Yes, AI threatens the folks who think that knowing how to write JavaScript or C makes them smart and being smart is their entire identity.

I think people who make being smart their whole identity are a out to use AI to turn the rest of the population into indentured servants.

The comparison between ELIZA and LLMs is valid you boil it down to "humans evolved for 6-7 million years, had spoken language for 500k years, but have only had something non-human that could generate convincingly novel language well enough to hold a conversation for a few decades".

There's no inherent reason it can't turn out having a non-human generate convincing enough language for conversation isn't a complete evolutionary blindspot the same way the short form feed has pretty much one-shotted society...

LLM chatbots are software designed to manipulate, addict, and mine data. Just like social media before it. But anyone who bothers to read the output in a domain they understand will discover they aren't all that no matter who OpenAI steals research from.

The critiques of AI laid out it Computer power and human reason are mostly still valid.

Or perhaps their wallets. There are a lot of programmers here. Some of them think that AI is the death of their profession, rather than a change in it.

"It is difficult to get a man to understand something when his salary depends upon his not understanding it." - Upton Sinclair

I was a programmer. AI is the evolution of my career. I'm in an interesting place. I've never been good at interviewing and I have certainly only gotten worse due to an admitted atrophy of thinking in "code." But I've been putting more effort in designing systems (agentic and otherwise) and building things at work as for fun.

It's been an absolute boon to finally build out all of the fun side projects I had always dreamed of, and after showing one off to some people I might even be able to monetize.

On the other hand I acknowledge that other people dont want to embrace LLM driven development for one reason or another, and I respect that. People got into the industry for different reasons , but code was always just a means to an ends for me.

I, uhm, think this idea would apply in both directions.

It's just a compiler barrier (signal fence), not a memory barrier that concerns the CPU. The CPU is free to reorder loads and stores.

Was the public key directory even operational in 2012? What about revocation checking?

I think processing that information is mandatory now, but probably was optional/largely unimplemented in 2012. But maybe I'm off by five years or so?

Revocation checking seems like a big gap, yeah. I’m not aware of any public revocation lists (but I’m also not super familiar with the industry), so I can only assume (hope?) that there are some shadowy but highly accurate databases that KYC providers are tapping into.

The CA public key I just got off my country’s website, they were kind enough to just publish it :)

I expect that UD2 stops instruction fetching (beyond the current block) and conversion to µops. A software interrupt or supervisor call should probably do neither because most of the time, these instructions eventually return and continue executing the next instruction.

An interrupt or SYSCALL instruction could do anything, which includes remapping or overwriting the memory location it returns to. So no, these instructions can't be prefetched in any case.

The leakage wouldn't be from training, but from other uses of Personal Data.

As far as I understand it, users can opt out from the training aspect, but they cannot stop their conversations (“User Content”) being used “[t]o improve and develop our Services and conduct research, for example to develop new features”.

If it actually exists, the protection would like go into the other direction: prevent playback of arbitrary MIDI files on the digital piano, instead of protecting the MIDI data against playback on other devices.

Public email and IRC archives/logs usually work just fine without Javascript.

ihtabs preserve iteration order and have performance competitive with swiss tables (while not requiring as high a load factor): https://github.com/vnmakarov/ihtab

That is what they claim but I'm very sceptical - not that they can preserve order but that they get similar performance.

As far as I understand it, the other efforts have not reported most of their findings to upstream developers, focusing on critical findings only.

This is understandable because upstream interactions at scale are difficult.

In the case of big projects like curl the interaction seems a bit more complete. E.g. There are some other blog posts about how the engagements and reviews worked which go decently beyond a pre-filtered dump of high severity CVE claims appearing out of the blue.