pishpash

It should have never been a cloud password manager play. It should be hardware device only, and tied to the device. One passkey on each hardware device.

Uh, if you set up a passkey it becomes the default almost always. Then on another device it'll pop up looking for it and look like login failure and you need to "try another way". That's alarming to anyone, not to mention seniors.

How many PM's have they got? Each one justifies his or her existence.

Reading this is like reading "this is like WiFi but for RJ11 nerds" while holding a dialup modem.

Lots of ways. Selling author signings, talks, authorized copies, subscriptions/merch, sponsorships. How do newfangled "content creators" fund their work? We already live in this world.

Maybe not all creative works are deserving of monopoly profits just by sitting on the ass in any case, and should stand on their own merits by producing downstream value that can be sold for whatever they can be sold for, by whoever puts in the work to deliver the value to the end user in a competitive manner. You know, open markets.

Attribution I can see. Consent or payment beyond market value, why? Just because you put in a billion hours to make a shitty $1 value output I should pay you a billion hours worth of labor?

If you buy the movie you most definitely have a copy that some machine made and you can keep it forever.

Does buying the book not count? Why in perpetuity? Even patents expire after 17 years.

Maybe a continuous evaluation process, why is there anything to train?

An cross of SoT (source of truth) and SotA (state of the art). SotA looks weird, admittedly.

The past cannot contain all information about the future, or there is some superdeterminism going on, by definition.

Which naturally begs the question whether the excess returns in finance (outside of services provided for liquidity matching, risk transformation, etc.) is not just exploiting inside information of one kind or another. If you have inside information, your time series forecasts are going to be excellent and no simulation is needed.

Unless disconnected from the world, or very thinly coupled, most processes become hard to predict in a generalizable way for this reason.

That's a bit of a vacuous statement. Stationarity makes sample statistics meaningful due to the LLN, sure, but almost nothing is stationary (and even if something is, there is no way to know anyway, you can only assume). If you condition on enough variables and do enough transformations you may get something seemingly stationary and therefore trivially predictable. But all the practical complexity is in that structure you need to specify. The true prediction "problem", when people work on such "problems", really is in the stuff besides the thing that you can just use averages to predict.

Maybe AI is deflationary, in the right hands.

No they don't. Are OpenAI and Anthropic charging at cost, or wish to? No.

More basically, production cost matters only if inference is priced at commodity prices. That's not what VC's signed up for, which is rent-seeking.

I don't think tens of hours are necessary. There was a very good public talk by a mathematician from first principles that even a child can understand, all the way up to the Riemann hypothesis, in 50 minutes. I wish it was online. It was the best talk of any subject I have ever heard.

Well, OpenAI was "not for profit" and still has "open" in the name. It's the age of transparent doublespeak.

It's the AI frontier ethos. "I get mine." That's the culture now.

Speaking of which, it happened before, too, in the 1970's with personal computer software.

Most textbooks are bad too. It's worth looking at some sample content to find the good ones.

The types who do ugly proofs, when they write code, produce spaghetti code. It's the same thought process going into how to approach something.

That's what it was, wasn't it? You checked out some files and that locked them against other changes, then when you were done you checked in.

What's likely to happen though, is human idiomatic writing will degrade to AI level and the two will converge. Just like nowadays it is harder to tell if some (human) non-native English speaker wrote some English text online whereas 20 years ago it was very obvious.

3dfx and nVidia even put Matrox out of business. The 1990's were a true competitive paradise up and down the stack, not like today.

Macintosh and SGI (+AIX, various Unix) were in fact a common combination used as desktop and backend server respectively in many 1990's scientific labs including biology labs.

That's some idealistic nostalgia. Software is generally poorly built today, and it's evidently not big enough a problem to fix.

Sometimes this is exaggerated. How do people view slide rules, telegraphs, and gramophones, or other historical artifacts from a further generation ago? They aren't that strange, and neither should rotary phone or film be to the current generation.

So far every single one has popped. The record isn't particularly good for "this time is different".

Aren't those completely obviously points?

They'll have stale knowledge though. It's like someone stopped interacting with society five years ago. Are they employable?