That's because EVs are perfect for taxis. Many short distances, never far from a charging point and time inbetween rides where you can recharge in the supermarket parking lot while getting a bite to eat. And in the city, with stop-and-go traffic, EV are much better at handling that than ICEs who hate that. Plus the lower cost-per-km-driven matters even more when it's a profit margin thing.

Those of us that maintain/maintained *BSD has absolutely been contributing back to it. And so has most of the companies that have used code derived from it.

Your lack of understanding of the licenses and their economics shows through. I think you've been hanging out with copyleft FUD providers, since you're repeating copyleft FUD. Companies will contribute changes back to BSD codebases - just not always *all* changes. But more changes than if they instead got a source license for something else, like QNX or Windows. As a former BSD developer, my clear impression is that since the formation of Free/Net/OpenBSD the vast majority of relevant proprietary changes have been offered back to the open source projects. And by "relevant" I mean "changes that had a meaningful chance of being merged back into the mainline project".

Relevant changes not being merged back (e.g. that FreeBSD hasn't accepted the changes from MacOS) has typically been due to lack of review capacity rather than the changes not being offered.

Because he had been taken in by rms' propaganda about how the GPL increases freedom, as most of us was before we thought really really hard about it. Many haven't thought hard enough about it yet to change their minds.

This hides the cost of the GPL. A more appropriate formulation is

"If you want to force companies to give back all changes they make to your software at the cost of having them unable to make use of your software and contribute back at all in cases where they could only give back some of their changes, use the GPL."

I was a FreeBSD developer back in the day. I have worked for companies that made changes to open source software, both as the person doing those changes and in other situations where my team utilized open sourced code and other internal teams made modifications to it. My experience is that companies want to give back anything that is not strategic, and the changes they make to BSD-licensed code bases (really, all code bases) are mostly utilitarian rather than strategic. They want to give back for several pragmatic reasons: It avoids maintenance costs for their changes, it is good PR, their engineers feel good about doing it, and it creates a good relationship to the upstream project, meaning that if they need help, they're likely to get it.

In my experience companies contribute more to BSD/MIT licensed code bases than to GPLed codebases. It is very seldom that it would make sense for a company to make a change to a codebase under GPL and give it back and it would not have made sense to give it back under BSD/MIT. It may be different from individuals. Some individuals believe companies deriving from BSD/MIT-licensed code bases and don't give all their changes back are exploitative, and thus those individuals avoid working on BSD/MIT-licensed codebases.

Of course, if we convince people that companies working on BSD/MIT codebases is typically not exploitative, then the individual part will also impove.

To be clear: you can still totally argue about qualia, sentience, things of that nature. What is not up for debate is that they work through unimaginably-complex chains of logical reasoning on superpositions of conceptual states and engage in forward-planning and metacognition. These are just demonstrable facts.

There is no 'sentiment' in an LLM or internal state equivalent to 'thinking' anything.

Yes, there is.

No, it's called "Evidencing The Fact That It's A Widespread Shorthand" (in specific, among Wikipedians, where it's in official terminology as well), a widespread usage that was falsely claimed to not exist.

(I'll also take this opportunity to note: rage about the use of shorthand on "copyvio", but not a word about "algos")

Shorthand should be something that other people actually use

Okay Felicia.

TL/DR: "copyvio" is the Wikipedia-standard shorthand for copyright violation, used both in direct speech and in official templates. If a person says "copyvio", they've probably spent a lot of time hanging out at Wikipedia, where copyright violation is a frequent topic of discussion.

I'm sorry if shorthand offends your sensibilities. Should I spell out all shorthand for you? Maybe make sure I only use emdashes instead of "-" while I'm at it?

Oh, also, as pointed out further downthread, any LLM with a training cutoff before the license change date is pretty much assured to be okay for use in a cleanroom implementation. Copyright holders cannot make retroactive copyright changes, and open source licenses generally explicitly grant rights in perpetuity anyway. The threat would be new code. But since the cleanroom side has never seen that new code (it was released after the training date), it remains untainted. And training cutoffs generally lag model releases by many months, so not only should all current models be in the clear, but most next-gen models as well (subject to cutoff date verification, and various kinds of nuance such as avoiding RAG, potential post-training contamination, etc - but there's easy ways to guarantee these aren't problems that are present)

First off, you can always just use a LLM with an open training dataset (OLMo, Pythia, etc) to prove that the data isn't in it, and have that handle the "clean" side of the equation.

Secondly, you are under the mistaken impression that a LLM that has been exposed (which, BTW, does not mean "memorized", or even any rough understandings of the workings - ask the LLM (ideally an abliterated one to avoid refusals) to reproduce the original code verbatim, it'll almost certainly fail badly) automatically fails in such a suit. It does not. There is no such thing as "homeopathic copyright contamination". Copyvio requires substantial similarity. LLMs will tend to have common algos memorized, but those are the very things that tend to be simultaneously not copyrightable.

The cleanroom is an evidentiary shield, not a pillar of copyright law. If you can prove that the cleanroom team had no access to the original, then you can prove that they didn't copy it, and potentially get summary judgement. If you can't prove that, that doesn't mean that it's not copyvio, it just means that you instead have to focus on substantial-similarity defenses. Access without substantial similiarity is not copyvio. If you read an open-source codebase, close it, and then later write a new program purely on your own that looks nothing like the original, you had "access", but still did not infringe. The same applies to LLMs - but even moreso, because humans are much better at learning from brief, limited exposures to content than LLMs are. A court will usually apply an "Abstraction-Filtration-Comparison" test - they break the code down into layers (source code, parameters, architecture, algos, etc) (abstraction), remove everything that copyright law does not protect (ideas and common algos, things where there's only a limited number of ways to possibly do it, etc) (filtration), and then compare the creative expression that remains to see if there exists sufficient similarity to count as infringement. It is highly unlikely that a LLM-implemented version will fail that.

(That said, you're talking about a long, expensive legal battle if you get to that point, so you obviously prefer to be able to get summary judgement if you can)

What the GP described is known as a cleanroom implementation, and it's perfectly legal, and has been done many times in ways that have been upheld by the courts. The dividing line, by and large, is that the documentation has to be flexible enough that there's multiple potential ways it could be implemented, that you're not just transforming a code flow into "text pseudocode" and back. It doesn't matter if, by chance, it looks like the original - only that there were multiple means by which it could have been done under the specification produced.

There's also some various process-specific factors you need to manage - full separation of the teardown and cleanroom-implementation teams, a judge filtering all communications (with an audit trail) back and forth between the teams to make sure that they meet legal standards of separation, things of that nature. But a competent LLM can help set up the process and run each agent for each role in its own isolated context.

Why?

Seriously, I get you are personally outraged, but that's not a reason. Has he mismanaged funds? Is his strategy failing the company? Is he driving key persons away from NVidia? Has he sold out to competitors?

Which one, exactly?

predicted that it would be the greatest embarrassment in the history of automobiles.

Said about every single Tesla model and non-vehicle product, and about the company as a whole, every time, by millions of people, for its entire history. That's hardly a badge to pin on yourself unless you were also saying that the Roadster, Model S, Model X, Model 3, and Model Y, and grid batteries would be massive successes.

Also, don't confuse "putting down a refundable $100 reservation" with "prognosticating about the future". Re: prognostication about the future, that's not "putting down a refundable reservation on a product you can choose to buy or not based on its stats when it comes out", that's making investments in the market. And back in the day, I made a TON of money long on $TSLA, predicting the overall company's success. How much, perchance, did you make shorting it, like millions of people did, talking doom and gloom about how every vehicle model was going to be the greatest embarrassment in the history of automobiles?

There was plenty of interest in the beginning**, and the electric truck market really doesn't have good competition for a given price point. The problems were:

1) The price shot up heavily
2) The stats and capabilities went down
3) Musk went hard-right, and rightly or wrongly the Cybertruck became seen as a symbol of being MAGA.

These things together pushed away ~90% of prospective buyers. Combined with the fact that the Cybertruck's production line was unique, as well as being a generation back, and it made it hard for Tesla to justify owning it.

** - Including from me. Beyond the stats-for-the-price point as introduced, I liked the engineering behind the design. If you want to make it rugged, that implies a thick skin. So make it an exoskeleton to bear the internal loads. Want to maximize strength while minimizing the mass? Resist torsional loads with a braced cantilever design (the triangular rear). Which, with a retracting tonneau, creates a very long, gradual rear taper, allowing detached air (from the angles necessitated by an exoskeleton) ample time to retract and minimizing the drawn wake. On and on - from an engineering perspective, there was a lot of great stuff that appealed to me. But little of that ended up staying. The final design hangs the (thinner) exterior metal onto a frame just as dead mass. The big raft of changes led to mass growth, cost growth, and range decline (despite advancing battery tech and pricing). And of course, I'm not super eager for Elon to end up with any of my money these days. Another family member had also been interested, and also cancelled.