Musk, the World's First Trillionaire - Is He Worth It?

[Cyclotron]

I wonder if this comes down to the techbro belief that we can build self-improving models.  Obviously anyone who cracks that will have an effectively indefinite lead.  They also may well create Skynet (er, sorry, I mean "models not aligned with human survival").

Currently all AI models have to be trained by a huge amount of reinforcement learning from human feedback which is extremely expensive.  If a model can 'teach' itself (it needs to be self-critical) it might be capable of self improvement.  No one has shown a model can do that without degrading into slop.

Anthropic claims Claude is writing its own improvements but not weights yet. I assume this will happen as the models become more and more proficient these roles. The recently wild and possibly marketing driven frenzy around coding analysis is an example of the model's evolution into better code modeling in the network.

The models will likely derive some of our worst attributes. They are, after all trained on our communications and patterns. These things have already shown up in some of the less restrained models.

[tom66]

I wonder if this comes down to the techbro belief that we can build self-improving models.  Obviously anyone who cracks that will have an effectively indefinite lead.  They also may well create Skynet (er, sorry, I mean "models not aligned with human survival").

Currently all AI models have to be trained by a huge amount of reinforcement learning from human feedback which is extremely expensive.  If a model can 'teach' itself (it needs to be self-critical) it might be capable of self improvement.  No one has shown a model can do that without degrading into slop.

Anthropic claims Claude is writing its own improvements but not weights yet. I assume this will happen as the models become more and more proficient these roles. The recently wild and possibly marketing driven frenzy around coding analysis is an example of the model's evolution into better code modeling in the network.

The models will likely derive some of our worst attributes. They are, after all trained on our communications and patterns. These things have already shown up in some of the less restrained models.

It is writing the software infrastructure around the fundamental model, but AFAIK it isn't involved at all in the model design or training stages.  Of course, the AI labs are increasingly secretive about what they are doing.

It is certainly an arms race on AI right now, imagine if an unfriendly state got access to a model similar to Fable. 

[Psi]

It is certainly an arms race on AI right now, imagine if an unfriendly state got access to a model similar to Fable.

oh. I'm sure they already have.
Governments always build secret supercomputers better than the public ones. AI is no different.

[tom66]

The big concern is data storage, NSA is archiving tons of content in the hope that one day they will be able to break the crypto with a quantum computer.  Many existing algorithms are NOT quantum-secure; it is absolutely possible to build ones that are secure but it hasn't happened yet.  Of course it may well be decades until we have viable quantum computers that can solve Shor's algorithm on large prime factors, or similar problems, but progress seems to have sped up in the last few years. 

[Siwastaja]

Anthropic claims Claude is writing its own improvements but not weights yet.

Reinforcement learning with AI feedback is a real thing, has been for a long time, so yes, AI trains AI; weights result in from a lot of autonomous AI work.

It's just that it has not completely replaced reinforcement learning with human feedback (and other steps with human participation, like, corpus selection/tuning, model architecture design etc.), and probably will not in near future.

[brucehoult]

The huge downside is that the majority of companies Musk owns shares in receive a crazy amount of government subsidies. First they blackmail local governments claiming if they are not given land and huge tax breaks they will build elsewhere. The billionaires building all these data centers are doing the same thing. The taxpayers are footing the bill and subsidizing these companies. Next they maximize government grants and subsidies which in some cases amount to 50% of profits. Again the taxpayer is subsidizing these companies. As such, it may be the biggest scam and transfer of wealth from the poorest taxpayers to the richest who pay no tax in history.

Surely all this is the fault of the politicians who try to tilt the market playing field by instituting subsidies, who offer tax incentives (and other things) to build in their district etc?

They don't HAVE TO offer those things. I personally would certainly prefer if they didn't — and I would prefer if politicians didn't even have the power to do that.

Given that they are offered, why would you not take the offers? If you don't then your competitors will, and then that's not a level playing field at all.

[tszaboo]

Cursor doesn't even have particularly good AI models... they're mostly an AI tooling company, wrapping around models like Claude... Hard to see how it's worth $60bn but the market stopped making sense a decade ago.   

It may have just been because Cursor had something working already. And it saves SpaceX time to simply acquire it now and continue working on that, rather than a 6-12 month delay to build their own solution while others get further ahead.

But $60bn to save 6-12 months is crazy.

It's also possible they have something new that we have not seen yet but SpaceX has.

I mean the headline figure is likely to be 60B.
Which is likely to be paid with their own stocks.
So you pay for an -over-inflated stock price company which doesn't make profit- with the stock of an -over-inflated stock price company which doesn't make profit-
It's really like buying toy houses with monopoly money, isn't it?

[temperance]

And while you are all discussing if Musk is Worth It you should take into consideration where you are in this.

From the wiki:
In his most recent work, Survival of the Richest: Escape Fantasies of the Tech Billionaires (2022), Rushkoff explored the calculus some of the extremely wealthy make in the recognition that their often single-minded pursuit of greater profits and better technology are creating an increasingly unstable world. In a 2022 talk for House of SpeakEasy's Seriously Entertaining program,[66] Rushkoff explained the billionaires' mindset as coming down to this essential question: "How much money and technology do I need to insulate myself from the reality I'm creating by earning money and using technology in this way?" He argues that treating people better in the present may be the most surefire way to avoid widespread catastrophe in the future.

And all this on behalf of everyone. Accordintg to Musk we must have a plan B in place if catastrophic events on earth take place. For who? Certainly not you but a class of people detroying the earth at high speed in their persuit for more money and power?

When they asked Musk why he doesn't contribute in a meaningful way to "repair" the problems we have here he seems to have answered: we should do both. I can't wait for his second plan to repair the problems we have here to unfold. (His electric cars are an illusion.)

Is He Worth It? No, not at all.

https://rushkoff.substack.com/p/you-are-not-crazy

[Psi]

His electric cars are an illusion.

You completely lost us with that factually incorrect statement.
They're not an illusion, they are quite real, and they work.
Lots of people all around the world can put solar panels on their house and run their car off the sun for free instead of burning fossil fuels and contributing to climate change.  That's a good contribution to addressing a meaningful problem.

No one person will solve the problem in its entirety. It's all small wins done by many people that help bring about the solution.  The EV revolution is part of that, there will be many others. Like ways to generate cleaner energy that EVs can use.  It doesn't matter that some people may currently be charging their EV from coal power stations, progress is always small and in little chunks. Cars that run off electricity is the first step.

IMHO you can only be critical of someone and say they should be doing more if what they are doing is less than what you are doing.

[tom66]

Yes, let's not forget that before Tesla, the auto industry's approach was very half-hearted.  We had cars like the 24kWh LEAF which could go about 80 miles with a tailwind, and the Chevy Volt which whilst innovative for its time was only a plug in hybrid, still ultimately quite dependent upon fuel.  Tesla showed that not only could you build a compelling electric vehicle, if you did so people would queue up to buy it.  They built the refuelling network for their cars which gave them a massive leap over the competition and showed EVs were viable for long distance travel. 

None of this means I like Musk... but I appreciate he was instrumental in the development of the electric car.  I would almost certainly not be driving two EVs today without him (even though neither of them are Tesla).  The biggest problem I have with Tesla is they seemed to become single-minded on eliminating some quite useful things in cars, like door handles, gear selectors and levers for indicators and wipers.  To save a few bucks they eliminate parts that can be replaced with software, which makes the car actively worse to use.

[Psi]

Yes, let's not forget that before Tesla, the auto industry's approach was very half-hearted.  We had cars like the 24kWh LEAF which could go about 80 miles with a tailwind, and the Chevy Volt which whilst innovative for its time was only a plug in hybrid, still ultimately quite dependent upon fuel.  Tesla showed that not only could you build a compelling electric vehicle, if you did so people would queue up to buy it.  They built the refuelling network for their cars which gave them a massive leap over the competition and showed EVs were viable for long distance travel. 

None of this means I like Musk... but I appreciate he was instrumental in the development of the electric car.  I would almost certainly not be driving two EVs today without him (even though neither of them are Tesla).  The biggest problem I have with Tesla is they seemed to become single-minded on eliminating some quite useful things in cars, like door handles, gear selectors and levers for indicators and wipers.  To save a few bucks they eliminate parts that can be replaced with software, which makes the car actively worse to use.

Yes. Someone being an arsehole, or a liar, or just a personality type you don't like is all pretty irrelevant when you think about it.
The only real thing that matters is the change they make on the world, and if that's in a good direction or a bad direction. Sometimes you don't even know for sure if what you're doing is having a good affect until long after you've done it.
The reason we do things is the only thing we can truly control.  Musk, as far as I can see, is motivated by trying to make things better for others.

[tom66]

Yes. Someone being an arsehole, or a liar, or just a personality type you don't like is all pretty irrelevant when you think about it.
The only real thing that matters is the change they make on the world, and if that's in a good direction or a bad direction. Sometimes you don't even know for sure if what you're doing is having a good affect until long after you've done it.
The reason we do things is the only thing we can truly control.  Musk, as far as I can see, is motivated by trying to make things better for others.

Personally I think, given Musk is a trillionaire (yes, unrealised wealth but still wealth), there is a lot of good he could do with the money he has beyond EVs and rockets.  I'd have thought a $50bn donation to good causes wouldn't go amiss.  He wouldn't even need to do that much to be involved; he is, after all, good at finding the right people.  Find the right person, give them the resources they need, and let them solve real world issues like food poverty, untreatable diseases, access to medicine, vaccinations, etc.  For all the hate Gates got when he was at Microsoft, he seems to be genuinely doing good things with his foundation.

[Monkeh]

Musk, as far as I can see, is motivated by trying to make things better for others.

Except those for whom he tries to make things worse. But hey, if you don't fall into one of those groups everything is fine, right?

[Cyclotron]

Yes, let's not forget that before Tesla, the auto industry's approach was very half-hearted.  We had cars like the 24kWh LEAF which could go about 80 miles with a tailwind, and the Chevy Volt which whilst innovative for its time was only a plug in hybrid, still ultimately quite dependent upon fuel.  Tesla showed that not only could you build a compelling electric vehicle, if you did so people would queue up to buy it.  They built the refuelling network for their cars which gave them a massive leap over the competition and showed EVs were viable for long distance travel. 

None of this means I like Musk... but I appreciate he was instrumental in the development of the electric car.  I would almost certainly not be driving two EVs today without him (even though neither of them are Tesla).  The biggest problem I have with Tesla is they seemed to become single-minded on eliminating some quite useful things in cars, like door handles, gear selectors and levers for indicators and wipers.  To save a few bucks they eliminate parts that can be replaced with software, which makes the car actively worse to use.

Yes. Someone being an arsehole, or a liar, or just a personality type you don't like is all pretty irrelevant when you think about it.
The only real thing that matters is the change they make on the world, and if that's in a good direction or a bad direction. Sometimes you don't even know for sure if what you're doing is having a good affect until long after you've done it.
The reason we do things is the only thing we can truly control.  Musk, as far as I can see, is motivated by trying to make things better for others.

This is very much my view. I don't know if I "Like him" or not but his actions and at least I view his motivations are to make a better outcome for many.

The issue of claiming things and not being accurate is a huge problem for some. See Cybertruck earlier in this thread.
I also consider his inaccuracies of the past but for the most part, he isn't lying about what will happen. He's getting the date wrong.

Elon made these promises as well:

-Tesla would be profitable. He promised it from 2012-2019 it did happen
-Australia Battery system in 100 days or its free and scaling continues in support of the green energy grid in Australia
-He original promised high volume production of Tesla cars and lots of people didn't believe him but they achieved it.
-Similarly Musk made big promises on global charging networks but those are a reality today and its the reason other EVs are successful.  Without that infrastructure people would not be as willing to buy the competitive EVs out there.
-At different points he promised reusable rockets that would land. The Dragon space craft, and Starlink that folks said wouldn't work.

We can stack those up against the CyberTruck production numbers miss, and the we ain't on Mars yet problem.

[Cyclotron]

Yes. Someone being an arsehole, or a liar, or just a personality type you don't like is all pretty irrelevant when you think about it.
The only real thing that matters is the change they make on the world, and if that's in a good direction or a bad direction. Sometimes you don't even know for sure if what you're doing is having a good affect until long after you've done it.
The reason we do things is the only thing we can truly control.  Musk, as far as I can see, is motivated by trying to make things better for others.

Personally I think, given Musk is a trillionaire (yes, unrealised wealth but still wealth), there is a lot of good he could do with the money he has beyond EVs and rockets.  I'd have thought a $50bn donation to good causes wouldn't go amiss.  He wouldn't even need to do that much to be involved; he is, after all, good at finding the right people.  Find the right person, give them the resources they need, and let them solve real world issues like food poverty, untreatable diseases, access to medicine, vaccinations, etc.  For all the hate Gates got when he was at Microsoft, he seems to be genuinely doing good things with his foundation.

Gate's use of his philanthropy for power is the substantial problem there.

These books have had a substantial impact on my views on Gate's and others philanthropy:
https://www.amazon.com/dp/0451493249
https://www.amazon.com/Bill-Gates-Problem-Reckoning-Billionaire/dp/1250850096
https://www.amazon.com/dp/1784786233

This is the problem I alluded to earlier in this thread with Gate's and others. They aren't "Just giving their fortune away" as they'd like you to believe.

[PlainName]

The issue of claiming things and not being accurate is a huge problem for some. See Cybertruck earlier in this thread.

Yes, let's see that: he promised something, managed to achieve not very much of it, but that's OK because LOOK AT THE CAT.

[Cyclotron]

The issue of claiming things and not being accurate is a huge problem for some. See Cybertruck earlier in this thread.

Yes, let's see that: he promised something, managed to achieve not very much of it, but that's OK because LOOK AT THE CAT.

Look everyone, PlainName is Gas lighting again. The standard distortion and implied name calling and mud slinging as usual. Anyone surprised? 

Musk really really hurt your feels over that CyberTruck is the best I can tell.
Keep the gas lighting coming, its funny.
 

[PlainName]

and implied name calling

So you've said more than once. Remind me what names I am supposed to have called you?

[Cyclotron]

and implied name calling

So you've said more than once. Remind me what names I am supposed to have called you?

Can you even stop gaslighting?

[PlainName]

and implied name calling

So you've said more than once. Remind me what names I am supposed to have called you?

Can you even stop gaslighting?

Are you sure you know what 'gaslighting' means? Asking you to say what names you think I have called you isn't it. But your not being able to substantiate your claim certainly is.

For reference, and if you are implying that I am making up that you allege I call you names, this is your comment alleging so (my emphasis):

You pick a nit, call me names, the same old thing, mix and repeat.

Would you like to either substantiate that or agree you are mistaken?

[Cyclotron]

and implied name calling

So you've said more than once. Remind me what names I am supposed to have called you?

Can you even stop gaslighting?

Are you sure you know what 'gaslighting' means? Asking you to say what names you think I have called you isn't it. But your not being able to substantiate your claim certainly is.

For reference, and if you are implying that I am making up that you allege I call you names, this is your comment alleging so (my emphasis):

You pick a nit, call me names, the same old thing, mix and repeat.

Would you like to either substantiate that or agree you are mistaken?

 
You know, gaslighting, like trying to twist facts that are directly comparable, but you imply my comparisons of them are less than intelligent.   

I recycle used coffee to strip paint. I figured 20l gallons/day is easily achieved. OK, hardly any demand but it is used far more than any competitor (at least a cupful every day here) so is actually brilliantly successful.

I don't dismiss it as a failure relative to Musk's statements. That's true. But an objective look at the market indicates it is the top seller.

What's the inverse of MDS?

Saying that coffee paint removal is the same as my comparisons is exactly gaslighting, twisting facts.
Implying that my factual statements indicate the inverse of MDS is name-calling. 

The issue of claiming things and not being accurate is a huge problem for some. See Cybertruck earlier in this thread.

Yes, let's see that: he promised something, managed to achieve not very much of it, but that's OK because LOOK AT THE CAT.

Again gas lighting by twisting facts of the comparisons and data points I presented. As if saying the Cybertruck is the 7th best-selling EV is somehow unrelated to its market success.

Continue to pick a nit and chase it to see if you can score a point for some game that is only being played in your head. I don't know why you feel the need to attack my opinions more than anyone else's. Other than a long-time obsession of yours to want to pick a fight with me. 

If you don't like my opinion or comparisons, then ignore them.

Now that I've substantiated my claim, implied name-calling and mudslinging, are you ready to apologize and move on?
 

[PlainName]

If you don't like my opinion or comparisons, then ignore them.

You know, that's always an option for you too. Maybe if I annoy you so much you should consider it.

[Zero999]

Ah yes, the Vegas Loop.  A public transport system that's:
- Less efficient than buses or trains
- Requires one human driver for every four passengers
- Achieves 20% of the capacity promised by Tesla
- Is still not self driving

Ah, yes, such a balanced view of the test.

I've been on this loop twice and will try it again in a month when I'm back. Seems pretty cool, but I also like the other cool options like the Amazon
Zoox and Google Waymo are coming, which I also like.

Here are some points to help you balance your overall view of it.

-Reduced Travel Times: Shuts down a 45-minute cross-campus walk to a 2-minute ride.
-Lower Construction Costs: Built at $52.5 million, vastly cheaper per mile than traditional subways.
-Proven Passenger Throughput: Successfully moved over 25,000 passengers per day during major conventions.

Electric bikes would have been far cheaper and more effective.

The downside is electric bikes are much more dangerous, especially given the fact they're all too often used by irresponsible people and children.

[Cyclotron]

If you don't like my opinion or comparisons, then ignore them.

You know, that's always an option for you too. Maybe if I annoy you so much you should consider it.

Didn't think I'd get that apology. Came after me to explain, and when I did, you just moved on. This is the standard. Gaslighting, sling some mud, claim you didn't, more gaslighting, mix repeat.

I didn't say you annoyed me. I'm honored that you keep chasing me. 
Where to next, pal?
 

[Cyclotron]

Ah yes, the Vegas Loop.  A public transport system that's:
- Less efficient than buses or trains
- Requires one human driver for every four passengers
- Achieves 20% of the capacity promised by Tesla
- Is still not self driving

Ah, yes, such a balanced view of the test.

I've been on this loop twice and will try it again in a month when I'm back. Seems pretty cool, but I also like the other cool options like the Amazon
Zoox and Google Waymo are coming, which I also like.

Here are some points to help you balance your overall view of it.

-Reduced Travel Times: Shuts down a 45-minute cross-campus walk to a 2-minute ride.
-Lower Construction Costs: Built at $52.5 million, vastly cheaper per mile than traditional subways.
-Proven Passenger Throughput: Successfully moved over 25,000 passengers per day during major conventions.

Electric bikes would have been far cheaper and more effective.

The downside is electric bikes are much more dangerous, especially given the fact they're all too often used by irresponsible people and children.

They are a blight on the scenery too and contaminate rivers and so on as folks toss them off bridges, etc. The worst is they take up valuable space along walking and bicycling routes in cities, and generally, those who seem to be interested in community by bike don't use them at all.

I have often thought about jumping on one when I leave a pub snickered up.  That would lead to your "dangerous" outcome.