Unfortunate title when this is actually a better explanation of what OpenAI actually did and did not with Navier-Stokes

I remember tables within tables within tables. Were those really accessible? The humans writing that html couldn't always keep it straight

Yes? What's problematic with nested tables? You just tab in and out of them.

Now try to navigate something like the WSJ/NYTimes home page.

If you remember the old days of good UI, there was this idea of "tab groups".

Light theme and dark theme?

I'd say mobile and desktop (from same html) would be more like css zen garden, than light/dark.

Not the same thing. The kindle ebook you "purchased" could have different text each time you open it up. A movie you stream could be different every time.

Reading your comment I wonder if the parent comment is different each time, because "does the ebook text change reading-to-reading" is not the question being discussed.

No need to decrypt https, the other end of your connection does that and hands the info over to governments

Yay cloudflare!

I don't know, people. We still really don't know how OpenAI or others are producing these results. It's all very hand wavy and trust-me-bro. How much money/time/compute have they really thrown at these problems? How much human involvement was there? What LLM did they even use? How much regular software was involved? They have given answers to some of those questions but no proof that that's actually what they did. I don't know if it's worth giving them this much credit (which is what we are doing by writing these essays and spending so much time debating). Anthropic wrote a C compiler that turned out to not really be a ready made replacement for GCC. Did they ever do any more work on it? Has anyone else produced a C compiler? It seems like that and these proofs are just demoware that are not (yet? Who knows?) production ready to turn the world upside down. Impressive one-off demos, yes, but companies have been pulling those off for centuries without ever going anywhere afterwards.

A very important point! In Tristan (NYU prof)’s write up he noted evidence of the OpenAI mathematician team doing a lot of correction and guidance along the way. We are never told about this with openness and clarity. At a minimum, complete disclosure and honesty is needed by the companies and about the precise role of their staff members.

The OpenAI staff kindly corrected the LLMs naive tries at solving NS? Impressive staff I must say.

They do have Fields medal winners among their staff.

Bread ties are good too. Sometimes they're too short, but you can twist a few together

There are similar long coated-wire ties used for gardening.

Zip-ties, velcro, and twine (all suggested previously) also work well.

Waymo (and rideshare) replace regular human drivers. No need to compare only against rideshare drivers.

Um, in those two examples where it was ridiculously not waymo's fault it was still absolutely the fault of a human driver. I'm struggling to imagine a significant amount of cases where it's ridiculously not the fault of any human driver.

Thats why he specifically called out CDL drivers, or in another part of this thread they talked about rideshare drivers. Certain segments of humans likely approach the same rates as the driverless cars.

So? The vast majority are not CDL or rideshare drivers.

What does "with no substantive human input" mean? All of the training data is human input, isn't it?

They also train on synthetically generated data.