Underground, by Murakami, was one of the better books I've read. It mostly covers victims, but the later chapters are interviews with various members of the Aum cult. It does a pretty decent job of presenting the psyche of the type of person who joins, stays, and some that leave. Although mostly relevant for what he calls the "Japanese psyche" I still found it very insightful.
I think this is a really solid move. This gives OSS contributors a lot of flexibility. You could set the limit to 0, and manually add contributors. You could set it to 1-3 to allow people to get their foot in the door. But the de facto limit today is infinite, which is spammed. Imagine if GMail did this! If I don't whitelist or reply within `n` emails, youre done. I would KILL for that.
This is a huge difference between NYC and SF. There are certainly some in NYC who would prefer Waymo (direct route, no driver chitchat), but I don't think many New Yorkers would be proud of taking Waymo's. Most people feel embarrassed to admit they took an Uber somewhere!
One thing I hate about GH's status page: sure "auth" gets to live under API, so API gets the downtime report today. But all of my remote git ops are failing, because i was dumb enough to use the `gh` cli to auth. Their status pages are not well defined, but they treat them as such, leading to inflated uptime numbers.
This makes it an instant non-starter for probably 95% of organizations. A lot of people are about to get in trouble for using it before realizing this.
Not a lot of discussion on this, but there is no way to turn off data retention for this model. IME this is the first time Anthropic has released a model without allowing you to opt out.
I loved playing trombone in school. It's such a simple mechanism, it invites a lot of curiousity, and this piece captures that well. Instruments like piano, violin, and guitar are very visual, essentially wysiwyg. Instruments like saxophones, clarinets, flutes, take a long to mentally map and reason through (this combination of keys achieves note X). Trumpets, and other 3 valve instruments map exactly to trombone positions! Eg. no-valves = 1st position, 1st valve is 3rd position, 1+2 is 4th position. But visually you don't see this, and it doesn't invite the curiousity. Trombone super unique in that you get a little wysiwyg, but you have to square that with embouchure. But learning trombone, and then mapping that knowledge to a euphonium, trumpet, tuba, etc, gives you a knowledge about that instrument (eg ok if note X is 1st valve, and note X+1 is 1+2, then i know adding 2-to-1 adds a half step, because position 3 is a half step from position 4).
My first ever EM showed me this piece ~10 years ago, and I still think about it a lot. One pattern I've adopted is to keep as much code to be synchronous as possible. On larger teams, especially when the slop-cannon is really going, I can at least depend on codeowners to tag me if someone tries to convert something to async (eg. adding a DB call somewhere), because they chain of things that need to be converted to async is so long. Then I can jump in and say "this entire chain of code is sync, if you want a DB call, do it somewhere else"
I could see this being true, except that other high end electrics (Porsche, Audi) have not sold well. So the theory would be:
> people want either $30,000 electric cars (Tesla), $100,000 electric cars (Tesla), or $500,000 electric cars (Ferrari)
I do think Ferrari is trying to expand their audience with the Luce, but not to Dubai housewives. Ferrari's are for Ferrari collectors. There exists the guy with a few already, who daily drives a Tesla. Probably hundreds of those guys! This is for them (IMO).
I interviewed at a startup, really early stage, who claimed to be on track for $100k ARR. We all know "ARR" is bullshit, but I didn't suspect the ~$10k monthly to be bullshit. It turns out this was $10k pre-discount, and this product was ~FREE for YC companies, who made up ~100% of this companies customer base. So revenue was $0, or very close to it.
This is one of those ideas I think makes perfect sense, but requires so much operational change for the entire stack, that it would be very difficult to scale:
- Convincing labs to run distributed, burst-y inference
- Convincing people to run their Mac all day, hoping to make a little profit
- Convincing users to trust a distributed network of un-trusted devices
I had a similar idea, pre-AI, just for compute in general. But solving even 1 of those 3 (swap AI lab for managed-compute-type-company, eg Supabase, Vercel) is nearly impossible.
> These are starting points only: customers are responsible for reviewing, modifying, and finalizing their own materials. Draft templates are not the same as “pre-filled evidence.”
Yeah, ok. BRB to start a bank where I template everyone a billion dollars, its up to you to be honest with how much money you have.
I can understand two 20 year olds committing fraud. I can't understand a team of engineers PRE-PUBLISHING A TRUST REPORT before a single field has been filled out. This is worse than fraud, its poor craftsmanship.