He's pointing out that labor has always opposed labor saving technology, despite that being the basis of our modern quality of life.

In the past, "labor saving technology" has always spawned alternate jobs that people could take with some retraining. This time it might be truly different. If one day AI can actually do all knowledge work, there might not be anything left for former knowledge workers to do. There's no physical law that says new technology necessarily produces 1:1 new, different jobs.

> In the past, "labor saving technology" has always spawned alternate jobs that people could take with some retraining.

Labor saving technology does not create enough alternative jobs to employ all those that it displaced, otherwise it wouldn't be labor saving.

Instead, the surplus created by these technologies allows that society to deploy labor on less immediately necessary jobs. These jobs weren't created by the technology, they were always there, but society did not have the resources to staff them (think education, research, academia, merchants, etc.)

This dynamic has been true since pre-historic times, so you'll need some extraordinary evidence if you want us to believe this time is different.

Most jobs for most of human history have not been "knowledge work" involving symbolic manipulation. Maybe all the marketers, business analysts and software engineers of the world can take up their true callings as plumbers, carpenters and dishwasher repair people.

You think that all knowledge workers of the world will accept their social and material downgrades without making wave? That they'll all be able to find manual work?

Who knows, maybe we'll come to value manual and caring work once AI can easily do all the moving-electrons-on-a-screen?

The financial and social hierarchy you allude to is not immutable. Programming was once a low-paid, low-status job done largely by women. It's only relatively recently that it's become a lucrative, high-status masculine-coded career.

The whole point of this project is to have an LLM that speaks European Portuguese, not Brazilian Portuguese.

Right, and my point is that if you use 80% Brazilian Portuguese during base model training + 20% European Portuguese as post-training, you pretty much get exactly that, except with a ton more of available training data.

What's your evidence for that?

And if the first 80% doesn't bias the language after post-training (which I think is what you're claiming) why not go for English or a mixture of languages, which is essentially what they did by starting with EuroLLM?

Evidence? Not so much, I didn't realize I was defending a PhD thesis here.

I speak Spanish, and have talked with people who only speak Portuguese, either of the variants, and also talked with Portuguese people before how they see their language, comparing it with Brazilian Portuguese, and vice-versa. So basically based on vibes and experience.

> And if the first 80% doesn't bias the language after post-training (which I think is what you're claiming) why not go for English

I'm not sure how many languages you speak or encountered in the wild before, but some languages are VERY different from each other, some are a bit different and others are basically the same with some differences. Doing what I describe for languages that are similar is easier than languages that are very different, for what I hope are obvious reasons.

> I'm not sure how many languages you speak or encountered in the wild before, but some languages are VERY different from each other, some are a bit different and others are basically the same with some differences.

I'm a dual citizen of Portugal and Brazil and I live in the US now, so that's my linguistic background. (Also studied bits of French, Russian, Latin and Greek.)

> Doing what I describe for languages that are similar is easier than languages that are very different, for what I hope are obvious reasons.

Not only are your reasons not obvious, your conclusion is actually wrong.

If the goal is to create an LLM with minimal Brazilian Portuguese bias (which was one of their main goals), it might actually make more sense to train it in any other language BUT Brazilian Portuguese (say, English), then fine-tune it for European Portuguese.

LLM's have shown to be very good at generalizing across languages (the transformer architecture literally comes from work on translators IIRC).

> If the goal is to create an LLM with minimal Brazilian Portuguese bias (which was one of their main goals)

Oh, I wasn't aware that was their goal, would certainly be intuitive to avoid Brazilian Portuguese if that's the case, although I'm still not sure it actually makes sense to 100% avoid it for pre-training even if you're trying to avoid Brazilian bias, you can "skew" things pretty heavily in post-training if you so wish.

Where can I read more about this goal, because it doesn't seem to be mentioned in the submission article, just a short off-hand about one of the benchmarks, so I'm guessing there is some resource they talk more about the specifically perhaps?

> Why would you let America take 2-3% of your transaction volumes?

I don't think VISA/Mastercard takes such a fee? (They'd be some of the biggest companies in the world if they did.)

The fees they charge are actually fractions of a percent, the rest are charged by the card issuer, which is usually your bank.

You could, in theory, use the VISA network and not pay those fees to a card issuer.

Still greater than 0.

There's absolutely no reason for a country to outsource paynent infrastructure to US corporations.

> You could, in theory, use the VISA network and not pay those fees to a card issuer.

You can not. The only way is to have a private agreement with the card issuer. That's why stores all try to push their co-branded cards.

Presumably they're making the leap that this printing technology can be leveraged to develop an alternative display technology that would change the structure in real time, kind of like color e-ink displays.

It's quite the leap, but that's science communication for you!

IMOD’s (like Mirasol, which was abandoned) are pretty neat:

https://en.wikipedia.org/wiki/Interferometric_modulator_disp...

https://goodereader.com/blog/electronic-readers/the-rise-and...

Especially since the breakthrough was to encapsulate the colored bits to keep them from clumping.

Are the quake, goldsource, source, etc engines jank? Because they all did this, to some extent.

It might be harder to find an engine from the early 2010's that didn't tie at least some physics to FPS.

Your examples are from 10 or more years earlier that Skyrim's creation engine.

That is to say, the older engines could have been limited by hardware requirements, or maybe decoupling physics from fps is an innovation that appeared between 2004 and 2011. Or maybe they are also jank.

Notably, the source 2 engine (2015) decoupled physics from fps (as I understand).

I'd say Skyrim was just around the border where you started to get engines which uncoupled physics from FPS.

This only really matters for competitive games though, so it's not surprising it wasn't prioritized.

> to some extent

You answer your own question! Not remotely to the same extent. Quake, for instance, gave a small advantage for jump height [with high FPS]. Skyrim would outright break.

Also, you've listed three generations of the same family; goldsrc, the child of Quake, predates Skyrim/Creation by at least a decade. Of those, Source would be the timely match. Not even close to the same amount of jank. Just... no. You aren't tricking me into writing lists.

I don't really intend to be critical of Skyrim, like many: I love the jank. It's expected. It's a Bethesda game.

Considering two Germans in Paris independently discovered the transistor just a few months after Shockley's team, this seems like a self-serving fantasy: https://www.computerhistory.org/siliconengine/the-european-t...

There's no question in my mind that American industry and capital markets were far better at pivoting to this new industry though.

If you have a diode, then the transistor is only a small step away.

Semiconductor physics books require you to work through a lot of material until you understand the diode, and then the bipolar transistor is just one next chapter.

> If you have a diode, then the transistor is only a small step away.

It is not. We've had semiconductor diodes since 1874, but it took many decades to develop the solid state physics to understand how they worked and how to extend them. Crucially, you need some understanding of quantum theory (energy levels, Fermi distributions, etc), which was not developed until the 20s and 30s.

Even after they had the physics down, Shockley still spent over a decade unsuccessfully trying to get a FET to work (due to trapped charges which were not understood until the 50s). This is partially why the experimentalists, Bardeen and Brattain, are quoted alongside Shockley as the inventors of the transistor, even though Shockley had come up with a lot of the theory years before.

And the other side would say that if we cut welfare spending we'd have a reasonable debt to GDP ratio.

It's very much political and it's a joke to pretend otherwise.

This is kind of the point though.

Cutting welfare spending will get us no where. The majority components of the federal budget are Defense, SS/Medicare/Medicaid, and debt payments. Not the forestry service or what we commonly know as welfare. At this point, even cutting everything else to zero still lands us in deficit. (Unless taxes are raised.)

To be serious, we need to talk about what cuts are to be made to SS/Medicare/Medicaid and the military. But no one wants to have that discussion. So we throw out meaningless issues like welfare and the forestry service. We quibble around at the extreme edges, never addressing the central problems. That's the essence of the politics being discussed. Those politics make the issue impossible to fix.

I honestly don't know why it's so hard? I'd be totally willing to countenance the necessary cuts to the sacred cow programs at this point. Why is everyone so opposed to it?

> To be serious, we need to talk about what cuts are to be made to SS/Medicare/Medicaid...

To be serious, we need to talk about the funding of it, not the cutting of it. If we raise the cap, it gets more funding.

If we increase Medicare taxes and then go to a single payer system, it could be funded as well.

There is zero reason to have for-profit health insurance.

> To be serious, we need to talk about what cuts are to be made to SS/Medicare/Medicaid and the military. But no one wants to have that discussion

The military-industrialist complex is a socialist jobs program.

Yes "If" spending is cut or "If" taxes are not cut then you might have a balance.

But just implement balanced budget goals. Accept at most a deficit of 1% in the budget or whatever. Allow for a deviation from this to do QE but require a more qualified majority and limit to 1 year only.

Want to cut taxes? Fine - but don't do it with deficit spending. Want to increase welfare spending? Fine - but remember to then cut somewhere else OR increase taxes.

The fact that one side can implement large tax cuts funded by borrowing over and over (and still be elected again) is absolutely _crazy_ on a scale that is perhaps only rivaled by the healthcare system.

They can say that, but it will not be true. The tax cuts, plus new military spending and new ICE spending dwarf welfare they want to cut.

Plus however many billions to bomb Iran and another 400 million for that stupid ballroom, the mad king's spending is out of control.

This.

ICE is now larger and more expensive than the entire United States Marine Corps.

Let that sink in.

Not only that, we also seem to start a new war every 6 months. Demanding money for each one of them. SS/Pensions/Medicare seem to trend nowhere but up. And like Santa Claus the party in power keeps handing out tax cuts.

We have to make a change guys. The old ways aren't working. We can't be distracting from the central problems by yelling "welfare!". That doesn't work anymore.

The only things that we can cut that will have any real effect on the deficit are Medicare/Medicaid, Social Security, and Defense.

You can see this here: https://www.pgpf.org/article/chart-pack-the-us-budget/

And if I didn't quit my job, I'd still be able to pay rent. Oh no, whatever is there to do in this situation?

You can play the twisty game but the fact is simple - if he didn't have the political capital to cut spending, then cutting taxes is irresponsible governance.

Nvidia subsidized machine learning research for years (both with CUDA, hardware donations and developing what was a very niche product line just for them) before deep learning became big, much less the advent of LLMs.

Certainly Jensen seemed to have an extremely long view on this burgeoning machine learning market in the early 2010's.

It didn’t hurt that they had a two companies named Intel and Microsoft that completely missed the boat where GPUs or mobile computing were concerned both are currently the top two companies in tech today by market cap?

The velocity factor is usually 0.6-0.7, never seen it as low as 0.5.

And it's set by the dielectric, not the conducting material.

No, the problem is getting a high power (hundreds of watts) and high uptime EUV source, there's no reason to think this is a step towards that at the moment.