Keep in mind these are dual socket systems, and that’s CPU without any GPU yet. So with the CPUs populated and a consumer-grade high end GPU added, those components are at 1500W, ignoring PSU inefficiencies and other components that can consume non-trivial power.

For USA, you almost never run a 20A circuit, most are 15A, but even then that’s considered short term consumption and if you run over a longer term it’s supposed to be 80%, so down to 1440W. Space heaters usually max out at 1400W in the USA when expected to plug into a standard outlet because of this. A die-hard enthusiast might figure out how to spread non-rendundant multiple PSUs across circuits, or have a rare 20A circuit run, but it’s going to be a very very small niche.

Ah, ok, that’s fair. I agree that codec/bitrate choice has made a lot of ostensibly ‘4k’ content look like crap, so why have 8k when many providers/internet connections won’t even cover the requisite detail to drive 4k in streaming.

Even if the electios are free in fair, I don’t think he’d be done in November.

The only way he’s “done” is if GOP loses every single last senate seat up for grabs. Every single one in Nebraska, Wyoming, Montana, Idaho, Kansas, Oklahoma, South Dakota, West Virginia, Florida, Texas, etc.

Hell, most analysts think that Democrats don’t have a realistic chance of even getting a simple senate majority, let alone a veto-proof one or even a filibuster proof one.

If they don’t then they cannot remove anyone from office, they cannot override vetos. Yes, they can decline to pass bills, but given their stance of ‘executive branch has supreme power’, they’ll just do illegal executive orders and ignore the courts unless the supreme court agrees with them. Trump is already declared immune from any and all crimes except by the Senate and has the ability to pardon any and all federal cases, and that’s assuming his own enforcement agencies even bother trying to punish anyone…

But then there are the differences.

Let’s say the COVID vaccinne triggered a whole lot of new pharmarcies that specialized only in vacinnes. Still good news for Moderna. Except those new pharmacies can’t quite afford the vaccines they set up their business to work in. Moderna’s stock is so high though, that they can leverage that stock to get money to invest in those new pharmacies to give them money so they can buy the vacinnes.

Then the pandemic passes and those pharmacies have no business and fold and their market cap collapses to zero and Moderna spent a bunch of money they didn’t actually have on now worthless equity, and their revenue and perceived value drops back to pre-bubble levels. Except even lower because they incurred liabilities that they didn’t have pre-bubble.

For the crypto bubble, nVidia went out of their way to keep their financials out of it. But for AI they’ve been giving their biggest customers the money they need to buy nVidia’s product. Basically a cyclone of big top line numbers self-funded but enough to drive the markets wild for nVidia stock. The big players have likely already ensured billions of more secure assets that won’t pop as hard and so “why not?” to play with the extra ‘free’ money to see how big the numbers can go.

start focusing on TVs that actually last now…

That only makes their “people need to refresh their sets for our bottom line” even worse for them.

BTW, 30 years ago TVs were expensive and still failed. There was a viable TV repair industry because it was worth spending the money to repair and easier to repair.

Anecdotally, my Plasma and my LCDs have been more problem free than when my family had CRT TVs back in the day.

Umm… ok, but that’s not really related to this article…

Everyone ditching H265 in favor af AV1 universally doesn’t make TVs sell any more or any more expensive.

Certainly his use of LLM was stupidly egregious, but he found that even by those standards the math results underpinning the LLM were way off.

Yes, they connect by PCIe and thus the physical mismatch may be overcome, but they also are now drawing 15kw. More wattage than any circuit in my residential breaker box can handle.

Even if you did, there’s not even a whiff of driving circuitry for a video port, so your only application would be local models, and if the bubble bursts, well that would seem to indicate that use case would be not that popular.

No I would expect that these systems get rented out of sold to supercomputer concerns for super cheap if a bubble pop should occur.

The server boards would pretty much have to come with them. Also, and if those cpus go as high as 500W, and as a result a lot of homes might not have a powerful enough socket to power them. Even without GPUs, might need something like a dryer outlet to realistically power.

They have already leveraged their stock beyond their entire pre-AI market cap. There is no return to the old days now. If the AI boom goes bust on them, they have left themselves impossibly exposed. They will owe more money than they can possibly pay back.

They took what should have been a slam dunk to sell shovels for a gold rush with a nice fallback to previous viable business into an existential threat to their business.

A lot of the “AI” layoffs were using it as a plausible excuse for layoffs they wanted to do anyway. So I don’t anticipate a lot of them coming back.

In fact, if the models are ingesting this, they will get dumber because training on LLM output degrades things.

Challenge there being that seems to have proven elusive. It’s not too surprising, but trying to use machine learning for robotics is actually really hard.

Driving is much easier, training data with video, audio, and other sensor input complete with how the human manipulated steering and two pedals.

But direct human interaction with the environment is both much more complicated than three controls and is not instrumented. They are trying to build training data from remote operators, but it turns out we aren’t very good at controlling these things remotely anywhere close to acting directly. We are terrible teachers and there’s a fraction of the actionable data that other more successful models had to work with.

If an AI sees a video of someone doing something, it can make a similar video, but can’t model how that might map to what it would see as unrelated motor and hydraulic operation.

Note that Tesla was clearly a viable business, I don’t see the justification for it being 3 times the value of ford, gm, Toyota, and Honda all put together.

Generally people are not challenging the fundamental possibility of these as viable business, just that they don’t make sense at their valuations.

Though I’ll agree that open ai particularly should get some skepticism. To the extent that actionable business models might emerge, I don’t see openai actually in a position to be a big party of any of it. Microsoft and Anthropic seem to mostly own business revenue, ChatGPT is generally not even providing the models people select when they are able to choose.

OEM license revenue represents a tiny tiny bit of their financials these days. They could just charge nothing for it and business wise no one probably notice much of a difference.

It is foundational to a lot of what they do, but older devices are just as good for their subscription and tie in revenue. Hell I use my work subscription for office from Linux, complete with OneDrive filesystem synchronization. Microsoft gets all their money from my headcount even as I don’t even use Windows.

But that capex could bite them hard if revenue falls to follow from it. That’s pretty much the only exposure investors care about.

Familiar but with a difference in my case.

I’ve spent my entire career alternating between two experiences.

One is being grilled why I an delivering what I think should be done instead of what the executives told me to do.

The other is getting awards and promotions when it turns out that I was right and the customers loved it.

It happened to work for me to do it my way, though my executives have usually simultaneously rented the implication they don’t have good vision, they also know how to leverage my success for themselves. Particularly this most recent promotion has been stalled to reward better drones instead, but it’s looking like they have to pivot back to rewarding the folks the paying customers actually like instead of those that feed the executive egos.

It’s not just an abstract number, but leveraging it changes the value. If they hypothetically tried to leverage 2 trillion with of their stock, it wouldn’t be worth 2 trillion.

Of course the needle would probably barely move if they tried to leverage 50 billion.

Oh robot butlers would be huge, but no one is even vaguely close to that. All the demos have been remote controlled, demonstrating that can make dumb robots, but that’s not really new…

Yes, but it sounds like the kid jumped into the street without looking and physics were not on favor of getting from 17 mph to 0 in time.

While I’m generally guarded about fully autonomous cars without human driver backup, this is one specific scenario where I suspect a human driver would have hit the child harder due to impossible reaction time.

Robo taxis are a thing… but not much of a thing and it’s not Tesla.