s17n

The article's thinking is pretty confused - SpaceX's work for NASA has been a huge success, and that... shows that NASA is failing?

Criticizing NASA's continued support for the ULA seems valid though. But then, one could argue that the reason that the ULA is so bad is because NASA made the mistake of allowing themselves to rely on a single contractor, and that spending billions of dollars on no results is still worth it just to keep SpaceX on their toes.

Enforced correctness is great. Manual memory management isn't appropriate for most applications (games of course are an example of where it is!)

I have literally never used a doubly-linked list in my life, and I'm pretty sure that most programmers can say the same thing.

As for the example... yeah, Python is pretty terrible (for writing production codebases, I think its a great language for short-lived, one-person projects). Interesting that you mention Python because if you're considering Python and Rust for the same use case that's pretty bonkers, for anything that you might possibly have used Python to do there are many more natural choices than Rust. If you wouldn't have done it in C/C++ ten years ago, you probably shouldn't be doing it in Rust today.

> Fetishization of Efficient Memory Representation: ... I don't understand what the point is here. Some people care about avoiding heap allocations? They're a tool just like anything else

The point is that dealing with the Rust borrow checker is a huge pain in the ass and for most Rust applications you would have been better off just using a garbage collected language.

The first question I have when I visit the site is, how is this different from Firebase?

He's suggesting that acquisitions aren't happening because of antitrust enforcement. I'm pretty sure that has no bearing on the type of tiny acquisition we're talking about here but then again I don't actually know what I'm talking about.

Weird that his example of YC's decline is a company putting together a hacky MVP based on open source code, which is an absolutely classic YC thing to do.

In my experience doing these interviews in Node and Python, it takes about 5 minutes to get running. If it takes more than that... maybe use a different codebase.

Yeah this part didn't make any sense to me - is he claiming that later interviewers will be heavily biased by talking to previous interviewers? If so, I think that's probably not as strong an effect as he's claiming, and its also easily correctable by just having interviewers not talk to each other.

I do agree about the "you only need 15 minutes" part, but the solution there would seem to be replacing four hour long interviews with four 15 minute interviews, rather than with 1 hour long interview.

Obviously, what people want is full time jobs. That's what the whole political battle over these platforms is about.

"A study from Australia and the United Kingdom by Lorraine Sheridan and David James[13] compared 128 self-defined victims of 'gang stalking' with a randomly selected group of 128 self-declared victims of stalking by an individual. All 128 'victims' of gang stalking were judged to be delusional, compared with only 5 victims of individual stalking."

more or less

Meetup has steadily gone to shit, and it's not hard to see why: you can't make money with a user-friendly platform for hosting community groups. Do you have a plan to avoid their fate?

Interesting, thanks! Not to get too far off on this tangent, but how is that different from the way VCs / investment funds work in general (taking fees from the LPs in exchange for their services)?

> They’re frequently syndicated at double-digit spreads

What does this mean?

I think most employees are mostly well aware that founders take a money off the table in every round, and I think that it absolutely does negatively affect morale.

The entire theoretical advantage of linked lists over vectors (constant time insertion and deletion) comes from not sequentially allocating them.

I feel like at Google, design docs suffer from being the main thing that goes in your promo packet, which leads to the writing being done more with promo committee in mind then the documents ostensible audience of people who are working on the system the document pertains to.

If it was a phishing email it would have a signal username, not just "use signal"

That'll show em.

I'm not talking about teenagers. If you think that your social life is the same with and without imessage you're wrong, regardless of how old you are.

Using any other app just adds friction - obviously your best friend isn't going to stop talking to you because of it but weaker relationships might not survive.

In the US you can actually lose relationships if you don't have imessage. None of the other apps matter.

Related to #2, professional investors understand that the implied valuation in a round with a >1x liquidation pref is, in a word, bullshit. As an employee, you too should keep this in mind when you are valuing your equity package.

Yeah, Uber used to just be what Uber Black is now, if you remember. I'm sure that both the consistency but moreso the average quality of Waymo will eventually decline (although the experience will of course always be more consistent than Uber). They aren't going to maintain a premium brand because they have billions of dollars of engineering costs that they need to amortize over the biggest possible pool of revenue.

That's because the cars are all brand new Jaguars that are impeccably maintained. If Waymo succeeds in becoming a widely used service that is price competitive with Uber, the experience won't be this good forever.

It's not fundamentally necessary, old-school predatory businesses like Columbia House were able to survive without an automatic way to collect on debts.

Lyft already had the ability to charge my credit card, as they do for almost everyone I assume. Robinhood actually just deducted the amount from my balance and then sent me a "you have negative cash in your account please add funds or we will sell positions to cover" message, which is a bit bonkers in itself.

I feel like the license definitely can supersede that - you obviously wouldn't be GPL compliant if you required everyone who got access to your code to sign an NDA, for example.

How can they restrict access to authorized developers but also use the MIT license? Can't any authorized developer simply post the code on the internet?