Jeff Bezos said it before, his business model depends on people not owning computers (https://www.thecooldown.com/green-business/jeff-bezos-artificial-intelligence-cloud-computing/), Sam just gives it an unoriginal an utterly creepy spin. I expect Musk and the rest will follow suit soon with similar comments about how necessary is for them to own and control networks(Musk), the vehicles (Musk), etc... It's all very anderseenish and the network state ...(or as old historians would say, serfdom)

Truth is all this tech-bro's are realizing that in order to keep all their money, they need to become techno-feudal lords: they have to own and control the means of all production so they can always partake and remain on top. It's the ultimate monopolistic impulse, become the utility from which all other business will depend.

Where the golden T's blackend the sky and oil is the air you breathe.

Azure can hardly be described as more than a bad beta of a cloud provider, with way too many issues to enumerate...

Github might become CrashHub

European here, lived in 3 different European countries and now in the US. And no, the costs of shedding employees is not the reason why there is a lack of innovation in Europe.

First of all the numbers they display are wildly hyperbole, so much they could be hypergolic propellant for rockets. True, there are higher costs of shedding employees, but not by the margin posted by the authors.

The actual reason, is far simpler: there is a substantially less appetite for risk from the banking and inversion side than there is in the USA: It's that simple. It is not a question about regulation, costs, simply put takes substantially more to raise even a fraction of the capital you can raise here, and that shows.

I think you missed me paraphrasing Archimedes

If you pay attention to how current LLM's work, you can clearly see they are equivalent to Markov chains (even though complex ones) but at the end they're statistical predictive systems without heuristic behavior. The only reason they simulate reasoning is due to the large size of the corpus they're trained upon. Give me a sufficiently large training set and there is no complex behavior that I cannot reproduce through statistical modeling

No, what the army of lawyers does is ensue that it is being done in a way that passes legal exam.

I have experience of how "volunteering" in Amazon works. It is neither compensated (as you are also expected to complete your normal tasks, so you will do over time) nor truly volunteer (or your performance review suffers).

Just because the threats are not explicitly stated, does not mean they are not there.

More and more big companies start to resemble the robbing baron empires of the 18 and 19th centuries, where basically you end having to pay for the privilege of working there. Amazon has been firing employees left and right, and now is short? Well, why not start by hiring employees back? Is such a revolutionary stray of thought? Oh wait, it might cut down on Bezos's botton line by a few millions ...

And that my friend is the amazon way, make money by profiteering from free work

We have come full circle, msdos old edit.com with it's basic TUI is now the latest windows feature ...

Yes, with their Solarians. Is a critique of both the increased levels of automation but also of understanding individual freedom as an absolute without a counterbalance (the logical progression is that eventually a single solarian would see his freedom affected with a single other human being existing in the same continent, then planet, solar system, galaxy, etc... until a single individual per universe is achieved)

This is just one step forward to automate ourselves into extinction ... as eventually wouldn't be more effective to be just a virtual asset in an autonomous self-sustained cloud than a human that requires space, food, resources and real friends ?

Spoken like a true cowardnazi ...

Have the decency of putting your money where your mouth is and post publicly.

I for once saw these problems coming, but starship as a space system still is in denial about is two fundamental issues that will doom it to fail.

- The proposed mechanism for recovering starship is too risky and has no fallbacks. If they can prove that mechazilla can capture safely a starship after a booster has crashed into it, that will go away. Otherwise, lose of mechazilla will mean stranding astronauts or letting them crash.

- Complexity. In its current form, it will require up to 10 refuels in orbit to be able to reach the Moon (and who knows how to reach Mars). This multiplies by more than 5x the amount of things that can go wrong.

The way I see it it starship is still 5 years away from being reliable.

The move towards governing by algorithm is very dangerous, specially as for one, current AI systems (LLM's) tend to hallucinate and are incredibly opaque to scrutiny of their decisions and choices. Other reasons include that there are a multitude of exceptions and special cases to deal with and while computation helps, a human should be ultimately accountable of the decision-making. The loss of accountability across the chain of decisions algorithms and/or AI systems will do would erode trust and open anything for a challenge in court.

Ultimately this also is largely undemocratic as congress and elected officials will lose visibility of the process...