Funny how people in this topic at the same time complain about systemd being multi-featured and then complain about abandonment of X11 in favor of of the more focused Wayland. :)

It seems to me that the real differentiator is not the number of features but software release date :)

Over here, far far away from USA, we have a joke story about a certain type of person, who finds a jinni lamp. Jinn tells that person that he can fulfill any single wish of his, but the caveat is that person's neighbor will get twice as much of the same thing. And the person responds with "gouge one of my eyes". And that's basically a tale of healthcare in some countries. "I will suffer some, but at least "those" people will suffer more!".

Because it is, for now. For a while at least. You can prove that LLM doesn't understand what it does and it is surprisingly simple. Request it to add two integers and then ask it to explain how it arrived at that result. The answer will be completely unrelated to the actual process LLM used because both results were generated independently and without understanding their meaning and connection.

This is likely true for the majority of humans too.

Objectively untrue. Any human who can add two integers or use a knife to cut food or write a word with a pen can afterwards describe what he did at least in some way. Unless he is lying which is a separate topic, we assume an honest attempt. If I wrote a word with a pen via execution motions with my hand, I wouldn't describe it as "I levitated the pen by manipulating gravity with my mind". Or if I added two integers, I wouldn't say that "I created a a lookup table of a many loosely adjacent numbers (different from the numbers in a task) and run statistical analysis on them and did a few more things like that a in a loop". No, I would say that I either calculated sums of decimals, or I I did a school technique with rounding up and then subtracted that adjustment later, or anything which actually happened. If I used a Python sum() in a CLI I would also say that I used exactly that not the other method. LLM can't do it.

Sure we do, since Fei-Fei Li and team created that annotated dataset, which allowed to train first LLMs. So LLMs are here for more than a decade already.

You are confused by what the L and L mean in LLM, or which data set she created, or both, or in general.

Or it is you who are confused. And I want to remind you that you can't retcon historical word use.

Fei Fei was annotating images... the second L in LLM is for "language". The first language models named LLM at the time were trained on language data, with an objective function of predicting the next token. It had nothing to do with the imagenet data. Imagenet data was used in... vision models.

The attention is all you need paper didn't ever use the term LLM or large language model because the phrase didn't exist in industry.

Why comment on a field you know nothing about?

Oh the joys of the free market. At least Google didn't have to suffer any red tape or taxes, and we all should deeply care about Google's suffering.

If they sell contracts on market prediction, then any person doing for example insider trading on corporations, fixing outcomes of sport events, directly paying out others to force event outcome, blackmailing people, or just directly adjusting some result like the CDG weather guy are completely in line with the platform rules and all that is perfectly legal. But since Poly and Kalshi themselves admit that those things are "unfair" and forbidden they have themselves removed their "just a contract" status and transitioned to a gambling status, where indeed fixing matches is bad.

Making a good progressive tax scale is a task which a freshmen can do in a hour. The problem is not a better tax scale, the problem is how to find actual shit to tax. Doing nothing, we have the modern state of economy, when there are poor people having small to no tax basis, then "middle(lol)" class exposing all their assets and funds fully to the tax office and paying maximum taxes on this planet among everyone because of that, and then passing middle class we have people becoming richer and richer but "somehow" paying fewer and fewer taxes, culminating in billionaires who legally have zero assets, zero saving and zero income. You tax tax such a person 5%, or 50% or over 9000%, the end math will always be zero. Good luck funding UBI from taxing billionaires :) . And salaried people actually paying max bracket taxes already can't fund the whole planet by themselves, math is not mathing here.

One thing that might help is closing loopholes around unrealized capital gains. Suppose that, as an alternative to taxing capital gains on its stock, a startup could deposit 20% of issued stock into a sovereign wealth fund? For a company with all its growth ahead of it, this is roughly equivalent and can’t be avoided later (not even with buy, borrow, die) because it’s already been paid. And yet, shareholders would probably find it attractive because they don’t pay the tax themselves.

Taking 1/5 of every startup from the beginning is a great way to get every entrepreneur to found their company in another country.

If other countries were that friendly to startups, people would already be doing it.

Startups get done in the US because of the network effect: this is where the other startups are, and where the people who fund startups are. It's hard for them all to de-camp at once to a different country, especially since most of the countries they'd want to go to already have higher taxes than the US.

There's probably some limit beyond which people would leave anyway -- presumably, lower than 100%. I don't know if they'd leave over 20%. But I bet you could demand a 5-10% stake, and get many threats but few actual departures.

Let’s say it’s voluntary and it only applies to whichever class of stock you designate, which can be non-voting. You use it for employee stock or to sell stock to investors who want a tax-free investment.

I think investors would pay more for the shares.

My total tax load is well above 50% if I include employer share he pays for me. That's a on very conservative bottom quartile IT salary. Far far from the USA ones. And then there is VAT which I won't even tray to include. I'm so thrilled to start paying 100%+ tax for OBI (Other's Basic Income) :) .

I didn't have such experience in my local polytechnic, but my friend went to study Theoretical Physics at the best university in the country. He told me, that right at the beginning of the very first day they gathered all freshmen in a hall and their old rector came to them. He said almost literally this: "here are a hundred of you, all wanting to learn physics, but we only need one". And eventually they did have that "one", he is a top researcher somewhere in America nowadays, he told me. And my friend finished masters there, completing all the hard study, got physics diploma, aaand... went to work as an accountant (he's CFO now, got a secondary MBA later). There is really not a high demand for theoretical physicists, even internationally :) .

Oh yeah, I can just imagine a line of people wanting to learn Strength of Materials course for fun, or anything equally crazy (requiring at the same time fresh knowledge of other hard uni courses as a prerequisite, a lot of strict rigor in studying and just pure baseline intellect to get it at all). One can learn Latin for fun, starting from literally zero and stop any arbitrary level too. One can't learn ANY hard STEP topic without doing that continuously on a progression. If god forbid you take a gap in studying STEM courses, in a decade you will have to restart from school program again :) .

PS: I get all the idealistic mulling about how universities should be these utopian centers for the voluntary knowledge study and collaboration and the only result should be merit based. But real world doesn't work like that. If a country wants to have professional chemists or say welders for example, it must force kids through a lot of boring and hard mandatory study for years. Humanity didn't invent anything better yet, sorry.