This isn't true. Transformer based language models can be trained for specialized tasks having nothing to do with chatbots. A real world example of this is ESMFold. While ChatGPT is trained on human language one could train up a model from plasma dynamics data using similar underlying technology in order to provide useful generalizations for prediction and manipulation of plasma.

So the hackers enter bids in the name of real truckers. Don't these truckers "close the loop" on the paperwork? They didn't bid on that load, so why are they picking it up?

Nobody knows what LLMs are fundamentally capable of doing. Personally I think it is nuts for people to speak of human thought as something special when trivial algorithms executed by no mind or computer have accomplished feats (e.g. flora and fauna of the planet) greatly exceeding the sum total of all human efforts.

For all anyone knows it is plausible to run LLMs in some sort of loop that rummages through latent space until it stumbles on useful solutions especially in situations where evaluation of objective functions are cheap.

Search is everything. The achievement of any goal requires determination of one or more subsets of possible actions that result in a desired outcome from a search space of all possible actions.

Can we please stop pretending that Elon Musk invented well... literally anything. Zip2, Paypal, Electric cars, rockets that both take off and land, solar power, battery power storage, tunnel boring, cybernetics, LEO communications satellites, AI/LLMs, self-driving cars/robotaxis, humanoid robots, flamethrowers, eugenics, etc. Every single thing he has been involved in is an idea that has been thought of before and published and often implemented before by other people. The hyperloop is not his idea. Aside from the many, many people who have thought of it before independently, Robert Goddard published plans for a vacuum tube train about 120 years ago.

So, whether or not a hyperloop could be practical, I really wish people would stop tying ideas like this to Musk. I am aware of course that there is an intersection between his fanboys and enthusiasts for such projects and also of course that his promotion of such things is one of the reasons he acquired some of those fanboys in the first place. The problem is that he ends up muddying the discussion. I would much rather discuss the practical aspects of the concept.

I am on the fence about it myself. On the one hand, in principle, it is a perfectly valid replacement for air travel, potentially not only far more efficient, but far faster. On the other hand, practical demonstrations of a hyperloop have not really materialized so far suggesting a great deal of difficulty in getting it to work in the real world. There are, of course, a lot of critics, but I have yet to see a debubunking of the concept with any significant technical rigor. From the summary:

I mean, that's mostly a load of garbage. "Intolerable noise and disruption"? I mean, has this Christian Wolmar ever been to a city? Not to mention, does it need to reach city centers? I've been to plenty of airports and a lot of them are very distinctly not in city centers, so why would a hyperloop terminal need to be? Although it also has much more technical feasibility than dropping an airport there.

Integrating with existing transport modes... It's a train. Isn't the guy supposed to be a train expert? Trains that carry people stop at train stations, where people then get off the train and either get on another train or proceed to cars in garages, taxi stands, or lots, shuttles, boats, planes, etc. That's how terminals work. Cargo generally would not need to be carried by such a high speed train. There might be special cargo, such as a car shuttle, so that people can drive in, park their car and have it waiting for them to board at the destination. Whether or not that is possible depends on some technical aspects of the train design, but it's not as if carrying cars is an existing mode of planes or most passenger trains.

As for passenger safety at such high speeds if something goes wrong... Well, if it's a non-catastrophic thing that goes wrong, the passengers generally will be fine. If it is catastrophic, they all die, probably instantly. Basically the same as with planes, minus the terrifying last few minutes.

As for the rest of the "doubts", there seems to be no practical reason that the capacity would be much of an issue compared to planes. As for energy costs, we don't know yet, but the whole idea involves almost frictionless travel with regenerative braking at the other end, with also an additional cost to maintain the vacuum (and possibly an energy cost for cryocooling superconductors, etc. depending on design), so the theory is that it should be low energy compared to jet planes.

As far as the economics of it not working, it is obviously way to early to determine that. As it stands, it is questionable whether the economics of air travel even work, so criticism of the unknown economics of this kind of rival for air travel seems questionable.

Basically, as far as I am concerned, time will tell for this, but I don't think we should pretend this is a settled question.

Wall Street mode: Party like it's 1929.

Here's where the summary goes wrong:

Artificial Intelligence is in fact many kinds of technologies. People conflate LLMs with the whole thing because its the first kind of AI that an average person with no technical knowledge could use after a fashion.

But nobody is going to design a new rocket engine in ChatGPT. They're going to use some other kind of AI that work on problems on processes that the average person can't even conceive of -- like design optimization where there are potentially hundreds of parameters to tweak. Some of the underlying technology may have similarities -- like "neural nets" , which are just collections of mathematical matrices that encoded likelihoods underneath, not realistic models of biological neural systems. It shouldn't be surprising that a collection of matrices containing parameters describing weighted relations between features should have a wide variety of applications. That's just math; it's just sexier to call it "AI".

... we had to blue box the chat lines.

Now get off my lawn!

The Oracle codebase is such a unintelligible mess the only way it can be improved is with the assistance of a superhuman AGI.

It must have found the Dilbert comic strip. And learned something about job security.

What kind of strawman is that? I never said anything about never having had an energy crisis or shortage of metals before. I said that we had never replaced ICE vehicles with EVs and had problems finding enough electricity to charge them either previously or now.

I am not sure what your point is here. As it is, plants are not as efficient at producing energy from sunlight as solar panels are, also corn is not the best energy crop anyway, and additionally making ethanol from an energy crop involves multiple steps that waste energy at each step. Might as well just burn the corn directly if you're using it to make electricity. Corn ethanol exists because there is a powerful corn lobby in the US that gets corn subsidized and also because, after the subsidies, ethanol is a cheap way to raise the octane rating of fuel (ethanol has an octane rating of 100-109, equal to or higher than iso-octane itself) which is useful for the petroleum industry in their "premium" gasoline scam.

They will put them at what they think is the optimal overlap of cheap land, electrical access and available data bandwidth. While they indeed don't care what they clear cut, the locations are still unlikely to be in forests. Also, aside from that, while they certainly are not environmentally sound, data centers are generally pretty compact in terms of land use, so this is pretty ridiculous. As for transmission loss, it is small enough that colocation of solar or wind power makes little sense for large, power hungry installations unless somehow the location is also perfect for solar or wind. It is actually more logistically sound to put solar and wind where they produce the most power and transmit that power.

In the US, the majority of household power is used for heating and cooling at around 52% with the second largest usage being water heating at about 18%, so those together account for 70% of usage. Now, to be clear, those are not all electrical in all homes. Obviously some homes use natural gas or even oil (sometimes wood) for heating and/or hot water though virtually never for refrigeration (technically you can have fossil fuel powered refrigerators like in many RVs, though even those are moving to battery and solar power). However, enough of that portion of household usage is electrical for it to still be a significant fraction of average household electrical usage. Refrigeration is around 5%, other appliances for cleaning, cooking, etc. take up somewhere in the neighborhood of 10% with lighting being around 7%. So the categories you mentioned (except for "central air" which falls under heating and cooling, which I covered, and of course EVs, which are also clearly covered in what I said before with typical miles driven per household) are only about 10% of household usage. So, aside from the TVs, game consoles, computers, cell phones, and unnecessarily "smart" devices, the other categories I mentioned are all devices that have a lot of room for improvement: heating and cooling can be significantly improved just by improving features like insulation and other passive heat control measures in many homes. Then the active heating/cooling measures have huge room for efficiency improvement as well, such as by replacing many existing systems with modern, high-efficiency heat pumps. Ditto for water heating. Heat pump-based water heaters can provide huge improvements. For refrigerators and other home appliances that I mentioned, there's clearly room for plenty of improvement in efficiency with much more efficient versions of refrigerators, washing machines, dishwashers, dryers available, and things like induction cooking, etc. available. For lighting, even though LED lighting has made a huge improvement and is nearly universal now, easily fifty percent of lighting power usage could be eliminated if homes had occupancy sensors installed to control lighting.
As for the TVs, game consoles, computers, cell phones, and unnecessary "smart" features of devices, IoT and all that, yes there is waste there. There is definitely room for improvement and I certainly agree that there are too many devices that simply never power off (although it's not like you actually need to go for the breakers when they plug into the wall or even into a power strip that has its own power switch - on that note I should point out that the "central air" is normally on a thermostat). Still, they're only a small percentage of overall usage and what room for improvement there is, if improved, would lower that as well.
Ultimately, just in the context of household EV usage, the extra power needed could be made up for by home efficiency improvements. So, the argument I made still stands. I should point out in addition though that household electricity usage is only about 38% of total electricity usage in the first place, and EV power would only be a fraction of that so, even if not offset by household efficiency improvements (which it certainly could be), it would not be a significant increase in electrical usage.

I can't even understand what you're trying to argue here. I mean, aside from positing a data center with power requirements at least an order of magnitude larger than any singular existing facility, making the bizarre assumption that it would operate solely using solar power, and making the other bizarre assumption that engineers designing a data center that runs only on solar power and battery storage would somehow forget that they would need to build out enough solar capacity to charge the batteries, what is the actual point?

Ultimately, aside from some general ranting and apparently subscribing to the nonsense anti-EV theory that somehow electrical power availability can't scale to support them, it is hard to discern what you are ultimately actually trying to say with all of this.

I disagree, this effectively outlaws chatbots.

This isn't merely semantics. Training is a radically different operation from programming. The pretraining of LLMs is simply throwing vast amounts of text at them. The model is able to generalize what was learned from its pretraing and apply that experience to its context. There is no explicit programming.

You claimed the chatbot was programmed to manipulate people. What did you mean by that? When I hear statements asserting such and such was programmed to do something it carries with it a presumption of intent on the part of programmer. Is this an unreasonable interpretation of your words? If this is not what you intended to say what did you intend?

AI is not executing anything. It is merely passing code to an interpreter which executes the code outside of the model.

Random examples of things AIs are able to do without having been programmed to do them:

- Language translation
- base64 decoding
- Solve simple ciphers
- Adding fractions
- Writing code in a variety of languages

Investment is a tricky one.

I'd say that learning how to learn is probably the single-most valuable part of any degree, and anything that has any business calling itself a degree will make this a key aspect. And that, alone, makes a degree a good investment, as most people simply don't know how. They don't know where to look, how to look, how to tell what's useful, how to connect disparate research into something that could be used in a specific application, etc.

The actual specifics tend to be less important, as degree courses are well-behind the cutting edge and are necessarily grossly simplified because it's still really only crude foundational knowledge at this point. Students at undergraduate level simply don't know enough to know the truly interesting stuff.

And this is where it gets tricky. Because an undergraduate 4-year degree is aimed at producing thinkers. Those who want to do just the truly depressingly stupid stuff can get away with the 2 year courses. You do 4 years if you are actually serious about understanding. And, in all honesty, very few companies want entry-level who are competent at the craft, they want people who are fast and mindless. Nobody puts in four years of network theory or (Valhalla forbid) statistics for the purpose of being mindless. Not unless the stats destroyed their brain - which, to be honest, does happen.

Humanities does not make things easier. There would be a LOT of benefit in technical documentation to be written by folk who had some sort of command of the language they were using. Half the time, I'd accept stuff written by people who are merely passing acquaintances of the language. Vague awareness of there being a language would sometimes be an improvement. But that requires that people take a 2x4 to the usual cultural bias that you cannot be good at STEM and arts at the same time. (It's a particularly odd cultural bias, too, given how much Leonardo is held in high esteem and how neoclassical universities are either top or near-top in every country.)

So, yes, I'll agree a lot of degrees are useless for gaining employment and a lot of degrees for actually doing the work, but the overlap between these two is vague at times.

Big deal. That's only one out of four.