computerphage

Well, they don't.

Relatedly, this is the reverse of how trees get their mass:

https://youtu.be/2KZb2_vcNTg

The Pijul website says their patch model is "based on a result from category theory" and they link to a paper from 2013. But wasn't Smalltalk invented in the 70s? It seems to me that either the paper wasn't novel or that Smalltalk used a different system. Why do you believe it's the same?

Do you mean that we clung to single threaded compute for too long? Like we spent too many transistors on just making one fast core because it's easy to write code to use one core?

It's not just a regional problem if it's happening in every region.

What about other years? 1980? 1970? I would love to see this data over time.

> homes today are substantially larger

Is there any availability of smaller homes in actual markets? If I'm a buyer in the market do I even have the option to buy a cheap smaller home?

I actually would think it's the other way around: ~10% of studies that sound like obvious nonsense and have p of ~0.049 are probably NOT going to replicate.

> They're wrong; full stop, no qualifications, no asterisk, no footnote, they are totally wrong on this and it is inevitable that their opinion will change.

This sentence takes the place of a reason. You could have offered a reason why you were right or a rational argument about the matter. Instead, you basically yelled "I'm right, they're wrong! I'm right, they're wrong!".

If you're interested to learn more about how MRI works. I recently found an excellent video on the topic. It's 42 min and quite technical, but that's what I like about it.

https://www.youtube.com/watch?v=TQegSF4ZiIQ

It will still definitely help some use cases:

> Identical files will always have the same hash and can more easily be moved from one torrent to another (when creating torrents) without having to re-hash anything.

"Epstein-Barr virus, or EBV, is one of the most common human viruses in the world. It spreads primarily through saliva. EBV can cause infectious mononucleosis, also called mono, and other illnesses. Most people will get infected with EBV in their lifetime and will not have any symptoms. Mono caused by EBV is most common among teens and adults."

https://www.cdc.gov/epstein-barr/index.html

What are Berlin Tables? Are they the same as Rainbow Tables?

I agree with you, but it is also worth considering how much computation had to go into producing those couple of gigabytes.

> Indeed, eight years before Orville and Wilbur Wright took their home-built flyer to the sandy dunes of Kitty Hawk, cranked up the engine, and took off into the history books, Lord Kelvin, the President of the Royal Society of England made a forceful declaration. "Heavier than air flying machines are impossible," said this very powerful man of science....Rumor has it Lord Kelvin was slightly in error.

https://www.nasa.gov/audience/formedia/speeches/fg_kitty_haw...

So, you think the median estimate (50% estimate) is in 25 Billion years?

From the post: "I also noticed that my 2015 prediction placed 10% to 50% in a 5 year range, and 50% to 90% in a 20 year range. AGI is a long-tailed event, and there’s a real possibility it’s never viable, but a 5-20 split is absurdly skewed. I’m adjusting accordingly."

Though, obviously, offerings in other (foreign) markets won't necessarily file S-1s because S-1 is a concept of the United States' Security and Exchange Commission (SEC).

This is what happened to Shannon in the story. Her level was 79 ng/ml, but she still had high calcium.

So, cap and trade but only for one type of consumption?

Who?

And you think this is a good idea?

It wasn't unprecedented.

In my experience, Cargo is much easier and more reliable than the systems I've used across Java/C++/Scala/Python/Go/JavaScript. It is such a pleasure to use.

Looks like you missed the sarcasm

How big is "not that big"? What do you think it is?

> tunnel vision when it comes to X-risks, which leads them to overfund X-risk-due-to-AI over something like, say, X-risk-due-to-asteroids.

Risks from natural events like asteroids are actually quite well understood and we have tight bounds that they aren't that risky per century. The natural risks cause area probably deserves more funding on a global level, but EAs are definitely thinking about it. It's literally the first section in 80k's intro article about X-risks: https://80000hours.org/articles/extinction-risk/

Honestly, I'm surprised. This doesn't look like getting cancelled for being edgy. There's no mob that I can see pushing to doxx him.

It looks to me more like the reporter decided that Scott's refusal to use his real last name was a weird request, not a legitimate security concern for his patients and himself. The reporter just doesn't get it.

It's not like it adds to the story. Every on the internet knows Scott as "Scott Alexander". I didn't even know it wasn't his real last name until today. It just seems so cruel to insist on doing this to Scott when I can't see any good reason to do it.

"Results show that the average risk of severe injury for a pedestrian struck by a vehicle reaches 10% at an impact speed of 16 mph... The average risk of death for a pedestrian reaches 10% at an impact speed of 23 mph" [1]

But, importantly, that's for human drivers, who are probably braking when, or at least after, the collision occurs. An AV that completely fails a detection wouldn't necessarily even try to stop.

[1] https://aaafoundation.org/impact-speed-pedestrians-risk-seve...

All three of these things have different types.

(1) cargo on the command line

(2) sublime text

(3) Rust analyzer

So the "move from" didn't really make sense to me. You wouldn't stop using cargo (rust's package manager) to use rust analyzer (a code analysis / code completion tool).

To find the missing constraints, you can try asking what's wrong with the list [1, 2].

You might imagine trying to have an argument with a person trying to sell you a list sorter. The salesman is just going on and on about how "but look, each element of the list I gave you is in the right order! So you should buy my very nice list sorter!". I would want to yell "but you didn't include any of the numbers from the input list! You just forgot about 7, 3, and 5."

So, another formal constraint might be something like this:

    for each element, i, in the input list, the element i must be in the output list.

Then, you might again try putting in [7,3,5] and get back a list like this: [1,2,3,4,5,6,7]. Again, imagine arguing with the pushy salesman. The salesman says "Look, you wanted a list with the numbers 7, 3, and 5, and you wanted a list in order, this is both of those things, so you have to pay me for my fine list-sorter now!" And again, I would want to object and say something like "but wait a second, I want ONLY the numbers that I gave you in the input list in the output list!". So, you end up with something like:

    for each element, j, of the output list, the element j must be in the input list.

Now, you try putting in [7,3,5] again. What's the worst thing that might come out the other end that's still following all of these 3 rules? Can you spot the next problem?