"There are only two ways to find good new music - listen to a lot of bad new music, or outsource your listening choices to someone else - and the second doesn't protect you against the first."

Outsourcing your listening choices can look like lots of different things: that friend who goes to lots of concerts and always has an amazing new band they've heard recently, radio DJs, algorithmic suggestions like Pandora or Spotify, the Billboard Top 100, your local bar's live band choices, the Grammy Awards, going to clubs where DJs play new music, etc - but ultimately they come down to the same thing, letting someone else decide what you listen to.

And while my pithy version mentions "bad new music", included in there is anything which is not "good new music", including lots of mediocre or inoffensive stuff which doesn't rise to the level of being "good".

I first thought about it in the context of music, as I was looking for new songs to choreograph to, but it's true of discovering any new products where the quality is a matter of taste or subjective assessment.

- Want to find new food you like? You either eat lots of weird foods, or you find someone (a friend, a food blogger, the NYT food reviews, your mum, anyone) to recommend you try something they've discovered.

- Want to read a good new book? Either pick up random books, most of which will be trash, until you find something you like, or find someone to filter down the books (a small bookshop which carefully curates its titles, a library's recommended reading list, the best sellers lists, Oprah's book club, etc).

- New TV shows? Watch many bad shows until you find a good one, or wait for recommendations or awards nights.

- Restaurants, clothing designers, shopping malls, Youtube channels, content creators, movies, directors, websites, etc - the story is the same.

The only places where this does not apply, is in contexts which have objective measures which can be used as filters: if you want a new monitor, you can go to any store and filter or sort the options they have by objective measures like "display size", "resolution", "response time", "weight", "connectivity" etc, and find new products which meet the criteria. This is still dependent on someone to go and collate the information about all the products, but you are not forced to try lots of incorrectly-sized monitors to find one which optimises your preferences. Similar for microcontrollers, CPUs, car trailers, light bulbs, etc.

But even things with objective measures often have subjective qualities which have to be assessed - you can filter laptops on weight, RAM, clock speed, and storage, but how it feels to hold, whether the keys have a nice feel, whether the machine overheats too quickly - so you're often back to the original observation on these matters too.

> With their ability to shapeshift and manipulate delicate objects, soft robots could work as medical implants, deliver drugs inside the body and help explore dangerous environments.

It's probably less than three clicks on most home router web UIs.

High quality decompilers have existed for a long time, and there's a lot more value in making a cleanroom implementation of Photoshop or Office than of Redis or Linux. Why go after such a small market?

I suspect the answer us that they don't believe it's legal, they just think that they can get away with it because they're less likely to get sued.

(I really suspect that they don't believe that at all, and it's all just a really good satire - after all, they blatantly called the company "EvilCorp" in Latin.)

Pretending that it's English-native is why there's unspoken incentives to sound more "native", and thus use these grammar-correcting tools.

Some of the intelligent comments on here come from people who learned English in recent months or years, rather than in childhood.

Their English isn't always fluent or well-structured. If they rely slightly more heavily on suggested-next-word tools or AI translations, is that a reason to exclude them from the conversation?

Conversely, many English learning resources for non-native speakers focus on strict formal language, similar to AI-generated text. Do we risk excluding people who have learned a style more formal than we're used to?

Sure, but I still don't know who they are, so I can't give their name over for either investigating the theft or reassigning the speeding/red light/parking fine.

- Fall back to old IOT SSID with ancient protocols

meant 2.4G and not WPA3.

Riddles are such a big part of the human experience that we have whole books of collections of them, and even a Batman villain named after them.

These days we have netcat/socat and others, but they're not reliably installed, while telnet used to be generally available because telnetting to another machine was more common.

These days, the answer would be to use a netcat variant. In the past, telnet was the best we could be confident would be there.

But there's the rub. If you found the code on Github, you would have seen the "simple licence" which required you to either give an attribution, release your code under a specific licence, seek an alternative licence, or perform some other appropriate action.

But if the LLM generates the code for you, you don't know the conditions of the "simple license" in order to follow them. So you are probably violating the conditions of the original license, but because someone can try to say "I didn't copy that code, I just generated some new code using an LLM", they try to ignore the fact that it's based on some other code in a Github somewhere.

https://en.wikipedia.org/wiki/Aaron_Swartz

We agree that that was its initial stated intention.

However, what we have seen in practice is that it has resulted in the owner-operators of those machines banding together to restrict access to the machines unless authors sign exploitative contracts assigning their rights to the operators (which they interpret as "getting permission").

From time to time, I would reflect on the fact that Microsoft and other commercial suppliers were getting paid for providing services to us, but I was expected to work for free.

> Humans generate perfect code on the first pass every time and it's only LLMs that introduce bad implementations.

That's not what the "anti-llm experts" are saying at all. If you think of LLMs as "bad first draft" machines, then you'll likely be successful in finding ways to use LLMs.

But that's not what is being sold. Atman and Amodei are not selling "this tool will make bad implementations that you can improve on". They are selling "this tool will replace your IT department". Calling out that the tool isn't capable of doing that is not pretending that humans are perfect by comparison.

... for NP-hard problems.

It says nothing about the difficulty of finding or checking solutions of polynomial ("P") or exponential ("EXPTIME") problems.

The classical joke is: (this variant from Brenan Keller[0])

A QA engineer walks into a bar.

- Orders a beer.

- Orders 0 beers.

- Orders 99999999999 beers.

- Orders a lizard.

- Orders -1 beers.

- Orders a ueicbksjdhd.

First real customer walks in and asks where the bathroom is.

The bar bursts into flames, killing everyone.

[0] https://xcancel.com/brenankeller/status/1068615953989087232?...