There is a new trend started by Stability and picked up by OpenAI that will provide base models for fine-tuning for each country/language/social group. Various groups are reacting to one-size-fits-all AI models.

This is an excellent article showing how AI models could impact communities effort to preserve their language.

> OpenAI's Whisper is another case study in Colonisation

https://blog.papareo.nz/whisper-is-another-case-study-in-colonisation/

And a positive one:

> How Iceland is using GPT-4 to preserve its language.

https://openai.com/customer-stories/government-of-iceland

When you got just 300k speakers of a language, you don't want the TTS and language model to make the new generation learn it wrong because the model didn't have good enough training data and made many mistakes. Kids are going to use AI in their own language, hence the risk of low quality responses impacting their small community even more.

The king of capitalism unironically claiming that increased productivity frees people up to do other things is hilarious. Like, I hope so, but it hasn't been the case for at least the last half century, and if we leave it up to people like him, it definitely won't.

Unironically, The Matrix. As an example of how BENIGN superintelligence will treat humanity.
As humanity will be completely useless after the singularity, lying in pods and being entertained in an illusion is probably the best what humans will be able to do.

Oh. Yeah. Fair enough. But then a smart enough AI wouldn't need to be given anything. It'd take over the businesses itself. As it is now, however, AI simply isn't good enough. Get back to me in five years.

Apparently RLHF makes the model less calibrated. So the more morality you put into it, the less you can rely on its confidence.

Don't worry, 10 years later AI will be better than us at everything so we all become its pupils. The hardest task for AGI will be to bring humans along. Think it was hard to bring AI to human level? You should see how hard it will be to bring humans to AI level.

Moon (2009) has some cool themes

But your supposition is not falsifiable, right? Under what circumstances would you believe a billionaire philanthropist is sincere?

There's a big difference between mopping a floor or even designing and programming a product and running a business.

Just look at Tesla. Brilliant inventor. Couldn't bring a product to market if his life depended on it.

I think people who complain about billionaires "doing good" are still complaining about it but it just doesn't come up as much. Also, if a billionaire is genuinely being good, they don't need to care what a minority thinks they can keep doing good and the results will speak for themselves.

After all, we know the billionaires not doing good will keep not doing it and will have the results speak for themselves anyway.

You mean the billions being hoarded in physical assets that the company needs to function?

Is there a free midjourney?

You think wrong. CEOs are the ones who make all the big decisions that can see the company flourish or bring it to ruins. Now about redistributing the excess. Already happens. The rich pay 80 percent of the government expenses. And of course they hire a ton of people as well both directly to work for them and indirectly by buying equipment for their business.

The real issue in America is the government's lack of efficiency. But I agree about AI.

I love the humor

So you think the singularity is after 2050 and that agi won’t happen in before the 40’s

interesting

the fun thing is in a few years you get find find if you were right !

yes I think you're right. I think the time that regulation could be effective has now passed. Perhaps AI is the great filter. if we get through it intact a bright future beckons.

I think it might be fast non responsive aligned.

This fool

As smart as he is, Gates is still a Boomer. He can't envision a society where people don't work and simply enjoy their lives.

Gates should donate billions to Stability AI !

Dall-e 2.5 is looking pretty good actually. Huge step up from the base version

This reminds me some talk about how if you are a billionaire who does something to help humanity, they get tons of shit for not doing enough (and i agree) but when you are a billionaire and you just hoard your wealth and do nothing, no one complains about you. To be clear i think it is absurd that we are still living in a system where it is possible for billionaires to exist.. i mean i understand how we got here. History and human nature and corruption and greed etc. But it is wild how successful the brainwashing of people is to just take the world today and everything in it as normal.

Not a movie per se but Detroit Become Human has a great story.

There are trillion parameters models that have been trained on vast data oceans. But it is expensive to expose the general public to them. However, large companies are a different story. We'll transition to swarm intelligence over a decade. At a certain point the data centers will reach a natural ceiling. But 4 bit quantized, optimized instances would have spread around the world.

I bought and read his book about the Information superhighways. He was so wrong about so many things. Also all you need is 640k RAM. I don't listen to him anymore.