I would refrain from calling it "unhackable". How do you prevent stolen credentials? And how would you prevent double-voting? You can ignore second votes, but they can still flood the blockchain and make it difficult to audit. Wouldn't it be vulnerable to DDOS attacks?

It does not need to build weapons. It can do a lot of damage by hacking and controlling computers and manipulating information. It may manipulate people into doing the damage. We are already relying unnervingly on information we receive from the internet, which we must somehow judge as reliable or "fake news". How do we know which information is true and which is false? This is not a far-away future scenario. Current policies about climate change are determined by what people "believe".

It might hack military systems and make a nuclear power believe that it is under nuclear attack. It can create enormous chaos by hacking and disrupting almost any computer. It doesn't even need to "want" to do any damage. You may ask your super-intelligent AI assistant to help you find money to pay your bills, and it may do so by hacking a bank, or by convincing an unsuspecting person to send you money, or by spreading rumors which cause a financial crash which benefits you.

Wait until the AI models determine that human brains are cheaper than GPUs.

Breaking backward compatibility was Drupal's fatal flaw. Not only are old modules incompatible with new versions, the core database needs a migration to go to the next version after D7, and Drupal does not provide this migration.

Now you have a reason to use it: Save 30 seconds off your commute.

...and in the process maximize the street surface used for car traffic, endanger and scare away cyclists and pedestrians from formerly quiet residential streets, etc...

The arstechnica comments point out that once one service goes down, the usage on the other services goes up and they fail too. It's interesting that they did not disclose the cause.

Also, hyperscalers use hardware more efficiently than organizations than self-host, therefore open models that are self-hosted would require more hardware, that Nvidia sells.

The summary links to the wrong place on Gates' blog. Here is the correct one. But the pages are all mixed together so that as you scroll down you get to a different topic and the url changes.

A good limit on AI development would be to allow it to grow only with renewable power, so that it does not accelerate climate change. If you scroll down Gates blog, you'll see his thoughts on climate change. But there is no mention of it in "The choices we make about AI" essays. As if the world's problems are unrelated to each other.

Why did you choose GrapheneOS vs other alternatives, and why is the alternative Android OS ecosystem so fragmented? There is GrapheneOS, LineageOS, e/OS, ...

This is quoted from the phys.org article. It wasn't added in the summary.

The article says something else: "they then cut the bindings off in order to scan the books more quickly. The printed book is destroyed in the process."

The word "Big" is not mentioned in either the salon or politico article. It was added by editor david. Flock is a private company, and most likely not worth a $billion. If "People are sick of tech billionaires trying to control our lives", a Flock uprising is a diversion. The real Big Tech is in everyone's phone, stalking everyone's browser, and who knows what else. Big Tech is carefully sly and does not make Flock's amateurish mistakes. Big Tech tricks people into paying for the spyware that spies on them, or giving away their privacy for free, in exchange for convenience.

You may be an exception. The books you mentioned are reference books. I am worried that reference books will disappear, because most people won't use them. Which means that we will be at the mercy of getting our information from AI.

Reference book publishers might publish their books through an LLM: you can ask questions but nobody gets the whole book.

LLMs are trained on "large, uncontrolled media collections". But they could be trained on curated media collections.

The humanoid classroom assistants that companies want to introduce to schools may be a sneaky way to educate LLMs and get paid for it, by putting them in real classrooms with humans. With disastrous consequences for the privacy of these students.

Things don't fall out of satellites. But the whole satellite could fall out of orbit and the sphere inside might have "hatched" out of the burning shell.