The general public hates AI music. Here are other things the general public hated over the years. 1920s: microphones. The idea that you could project your voice was lambasted by critics for being "manipulative" when they were first released. Today, microphones are pretty much a requirement. 1980s: synthesizers. Called "fake music" by millions. Detested by so many they were outright banned by the British Musician's union. Today, they're used in virtually every modern production. 1980s: sampling. Considered "not real creativity", since you were just playing someone else's music. Today, they're one of the most commonly used (and enjoyed) musical techniques in hip hop. More recently, auto-tune. Vocoders. Equalizers. Randomized MIDI arpeggiation, like Flume. Etc etc. Many of these are instrumental to music today (pun intended). Yet people hated them when they were first invented. To be clear... sound is what happens when you compress air waves. Music merely takes that one step further with rhythm. If you tell others that *their* compressed air waves suck, since they were produced by a slightly different machine than yours, you are retarded and should be shamed for it. AI is not coming. It is here. When a critical threshold of people realize this, it will become the dominant method by which almost all music - and really media in general - is produced. But that doesn't preclude artists from existing! It merely changes the technology they will use to create it, by making it easier. In a way, it allows all of us to become our own personal artists; to extend our ability, and share the vision in our hearts and minds more freely than through instruments that previously might have taken decades to learn. I say this as someone who has played piano for over 14 years and identify deeply with it; AI is far better than I am. This is not a destruction of the old ways. If anything, it is akin to pouring gasoline on the already raging fire of our taste, our culture, and our desires. Humans love art! We thrive in it! And what art is more powerful than taking the collective gestalt of human knowledge - of human will, perhaps - and concentrating it all into one point, one intelligence, who can help us create something *more* than we could before? The public sees AI as anti-human and destructive to the spirit of music. But I consider it uber-human; it has the will make music flourish. It will align with our inbuilt desire to create, share, and grow through technology. Many incredible artists the world would have never seen will now, with AI, be able to create things that spark joy in the hearts and minds of millions. All of us will be artists, in our own ways. Lowering the bar for music production will let us share emotions more easily; propagate cultural memes (see recent Mace Windu music); and ultimately become a more responsive collective society. I don't see that as a bad thing, and you shouldn't either!

Modern renaissance painting

From concept to reality. @SpaceX has successfully deployed Starlink V3 satellites into Earth’s orbit and made contact with them for the first time. At scale, one Starship carries 60 V3 satellites, the same network capacity as about 20 Falcon 9 launches.

In less than 24 months there will be a clip of a high profile CRO explaining this reality to a (much smaller) room of sales reps.

how to set the right reasoning effort level i get asked this a lot, and i learned that because the underlying mechanism is not obvious, a lot of people don’t have the right mental model to think about how to use the knob so a quick crash course here: reasoning effort level in mainstream LLMs today means a “ceiling”, not a “floor”! setting a high effort level does NOT mean every prompt you send will use a lot of thinking tokens. you can try this very easily - launch your agent in an empty directory, use high reasoning, and say “hi” - unless your harness is totally messed up, the agent will respond almost instantly how does that work? the most typical implementation is that in post-training, the models receive a penalty for how many thinking tokens they use. how strong the penalty is depends on the thinking effort level that’s set at low reasoning, the penalty is very strong, so any long thinking traces get punished into oblivion at high reasoning, the penalty is weak so longer thinking traces don’t get punished too much, and are allowed to survive more often as long as they get good results the most important thing to learn there is - thinking tokens is always a penalty, not a reward. there’s never a rule that says “at high reasoning i’ll punish you if you think too little” - this is because more thinking tokens always means higher cost which is always a bad thing and should not be rewarded so with that in mind, you should think of reasoning effort as “how much are you allowed to think”, not “how much do i want you to think” setting it at low means “for this task you are not allowed to think much”. you would want this when 1) the use case requires low latency; or 2) you already know the task does not require additional reasoning setting it at high means “you are allowed to think more when needed”. the model will still try to think as little as possible according to what its training data teaches it. you want this when 1) the use case allows slower response; and 2) you are not sure how much thinking is needed for this task for some models, “max” is special because it forces thinking mode to be on, which forbids “zero thinking”. i don’t have a full inventory for which models do this vs not. and i generally avoid this mode because i don’t think it’s a good idea to say “every request has to have thinking tokens no matter what” practically speaking, i mostly only use two modes: - medium, when i already know the task is well defined for example, implementing a spec planned by a highly intelligent model. i avoid low because 1) i rarely require low latency; 2) even when implementing well defined tasks sometimes the intermediate context is still ambiguous this is also my default for firstmate because highly ambiguous problems are typically handled by crewmates, and i do want firstmate to be fast - xhigh, whenever the task is not well defined yet. this is most often used for planning and investigative work. i do this to give enough room for the model to decide how much thinking it needs hope this is helpful!

We're adding support for AGENTS.md to Claude Code. Starting today in version 2.1.277, if there is no CLAUDE.md in a folder, Claude will check for and use AGENTS.md. You can toggle this behavior in /config.