> I read a paper yesterday where someone had used an LLM to read other papers and was claiming that this was doing science.
I'm not trying to be facetious or eye-poking here, I promise... But I have to ask: What was the result; did the LLM generate useful new knowledge at some quality bar?
At the same time, I do believe something like "Science is more than published papers; it also includes the process behind it, sometimes dryly described as merely 'the scientific method'. People sometimes forget other key ingredients, such as a willingness to doubt even highly-regarded fellow scientists, who might even be giants in their fields. Don't forget how it all starts with a creative spark of sorts, an inductive leap, followed by a commitment to design some workable experiment given the current technological and economic constraints. The ability to find patterns in the noise in some ways is the easiest part."
Still, I believe this claim: there is NO physics-based reason that says AI systems cannot someday cover every aspect of the quote above: doubting, creativity, induction, confidence, design, commitment, follow-through, pattern matching, iteration, and so on. I think question is probably "when", not "if" this will happen, but hopefully before we get there we ask "What happens when we reach AGI? ASI?" and "Do we really want that?".
There's no "physics-based" reason a rat couldn't cover all those aspects. That would truely make Jordan Peterson, the big rat, the worlds greatest visionary. I wouldn't count on it though.
You seem to be generating output files compliant with ISO 10303-214, this standard has been withdrawn and replaced by ISO 10303-242, edition 4 of 10303-242 has just been published.
You'll definitely want to memorize the password to the backup service that has the last copy of your password vault after a disaster. :P
> Writing your passwords down on paper is actually less crazy than it sounds
I agree that physical security can be incredibly useful against a lot of modern threats... but we can do better. I wish there was a dedicated password-keeper device format of:
* A small keyboard and screen
* The data encrypted at rest by one master password
* Only permits upload/download of the the encrypted file over USB. With some companion software, you just plug it into your computer, computer copies the encrypted file to somewhere on disk that gets regularly backed up, the disconnects and beeps to tell you it's done.
* Sturdy enough that any "Evil Maid" attack needs to be done by a professional rather than a conniving roommate or jilted partner.
> You'll definitely want to memorize the password to the backup service that has the last copy of your password vault after a disaster. :P
Why? Write it down. Perhaps leave multiple paper copies around with some trusted people, like your lawyer and a safe deposit box at your bank.
Your proposed device seems a bit complicated. You can get pretty far with a piece of paper and this protocol:
Construct your password from two parts. (1) random gibberish you write down on paper, (2) a 'correct horse battery staple'-style part that you memorise.
Btw, have you looked into Yubikeys? They are better than password storage, because they can store your private keys and do signing with them. The key never leaves the device. (They can also store passwords, I think.)
> Why? Write it down. Perhaps leave multiple paper copies around with some trusted people, like your lawyer and a safe deposit box at your bank.
Those people would then effectively have access to your nearly-current desktop/laptop data from anywhere, especially since they would have to know who you are which greatly simplifies guessing your username/email.
> You can get pretty far with a piece of paper
Password Papers (A) never get backed-up, meaning they'll be locked out of basically everything if the house burns down and (B) I've already tried getting relatives using them to adopt exactly such a fixed+variable combo scheme.
> Why? Write it down. Perhaps leave multiple paper copies around with some trusted people, like your lawyer and a safe deposit box at your bank.
Those people would then effectively have access to your nearly-current desktop/laptop data from anywhere, especially since they would have to know who you are which greatly simplifies guessing your username/email.
> You can get pretty far with a piece of paper
Password Papers (A) never get backed-up and (B) I've already tried getting relatives using them to adopt exactly such a fixed+variable combo scheme.
> Those people would then effectively have access to your nearly-current desktop/laptop data from anywhere, especially since they would have to know who you are which greatly simplifies guessing your username/email.
For secret answers like this I have Bitwarden generate a set of words that I put in. The words are actual English words, so the 'random gibberish' moniker wouldn't be correct.
But at least the answer doesn't match the question.
I've also learned to store the question, as some websites make you select the question before providing the answer. And my answers don't allude to what the original question was.
> I've also learned to store the question, as some websites make you select the question before providing the answer. And my answers don't allude to what the original question was.
I usually pick the first or default question. But yeah, that order might change.
Passwords in this style (passphrases) are also much easier to transcribe to devices that don't have or support password managers, or when sharing a password verbally or in writing.
Did you choose a KaiOS phone specifically because other current non-smart phones lack support for Bluetooth tethering (a.k.a. PAN profile)?
As far as I can tell, all currently available models from all manufacturers are based on some Unisoc platform and offer no indication of support for this feature in their manuals. Did you happen to come across any alternatives?
I'm not very keen on KaiOS given the ubiquitous advertising baked into it (which is apparently their business model).
Reading comments like yours makes me wonder what kind of mental model of the world some people are working with. Russia does not need HN comments to tell them where train stations are.
It’s highly likely that Russia already has detailed models of every stations in Ukraine. If they didn’t before the war, they do now. Mapping public infrastructure doesn’t require a lot of spying.
But you have to understand that information control during war requires a shift of mindset. It’s better to start controlling everything which could be used by the enemy even if they probably already have it than try to establish complex rules. It gives good habits to people.
Of course they know where the stations are. But they don’t necessarily know the precise local of all the underground tunnels, exits, mechanical rooms, equipment, etc. The underground network is far more complex than what the consumer map hanging on the wall in the station shows.
During the Cold War, Russia managed to map huge parts of the world, sometimes with higher quality and more accurate measurements than the countries themselves! Especially noteworthy considering that some of those countries (like the UK) were trying their hardest to prevent those sort of maps being made in the first place, yet the Russians ended up with better maps of the UK than UK themselves.
Considering that that happened decades ago, I'm guessing their (and others) capability of doing those sort of things have only improved, not gotten worse. But that's just me guessing.
I'd like to remind you that Russia is not the USSR. Surely the technology has significantly improved since, but some capabilities are definitely lost. One example is them not being able to build more strategic bombers.
This is why agencies don’t published detailed plans (only schematics) of train stations and airports. I learned this when working on a project for the New York subway in the early 2000s.
reply