Hermitian909

Having taken part in many build-vs-buy decisions, there are no easy answers. The overriding factors in most decision relate to the specifics of the business. Some very common reason for buy decisions:

- You don't actually know if this feature will be important long term

- You have many urgent needs, the opportunity cost of waiting for "build" to finish is too high

- Your company is, for various reasons, unable to attract and retain the kind of talent who could maintain a system of this complexity (I realize this sounds like "we're not willing to pay market reates", I promise it's more complicated than that)

> So it doesn't get changed or improved, and it can never be replaced. It is the incumbent that blocks progress. A technological quagmire we can't extricate ourselves from.

This describes almost every tool in my toolchain that's over a decade old, this is just what happens. If you want to kill terraform and replace it with something better the usual bar is that it must be 10x better. If that's something you think you could do I'd be (genuinely) excited to see it :)

These issues are mostly separate.

As a society, we need doctor's to make more money than delivery drivers because we expect them to be smart, competent, diligent, and extensively trained (sometimes for >10 years). You will not get enough doctors that meet this bar if you don't heavily incentivize them to sacrifice a huge portion of their life to achieve excellence in their craft.

Home prices are high mostly because our society has promised the demographic with the highest voter turnout (old folks and property owners) that their homes will fund their 10+ year retirement - a time when discretionary spending jumps dramatically. Anything that threatens these high home prices is shut down immediately by the voting portion of the public. As a result "A good salary" can't afford rent.

It's already happening. Pinterest had top of market pay but realized they could cut total comp substantially by agreeing to remote.

> “Want” gets redefined somewhere between the first and second to last paragraph

I think I get what the author is going for. English lacks words to describe the degree to which you want something paired with how seriously you're willing to pursue it. A word that means "I desire this enough to seriously work for it given the sacrifices required" would be really useful and I've found quite a few people pull the same semantic trick with want that the author does as a way to communicate it.

Certainly, trying to push this idea into my mentees' minds is something I've had to spend an unfortunately large amount of time doing given how intuitive it feels like it should be.

I think I've been unclear on the problem. Trust inside a single team is not something I've found to be a major problem, but between teams. Let's be more concrete:

I need to build trust between between 6 EMs spread over three different directors and 2 vps + some very senior engineers. I need to do this so that we can solve one of the larger problems the business is facing. Most of these people have no good reason to talk to each other (or me) regularly. Without trust, one or more of these people will likely torpedo the proposal in order to avoid tying themselves to other teams and people they don't know.

Pre-covid the solution to this problem was (mostly) chatting over lunch. You'd get the right people talking, seed the idea, and build up to a proposal that everyone could agree to drive forward. IME this whole style of consensus building is dead in a remote world because async communication is too low bandwidth for trust building.

This makes it really hard to take good organizational bets, you have to wait till you have a mountain of data and customer feedback before you can sell big projects. The quality of the company's output really noticeably declines from this.

Even well done, this kind of communication doesn't substitute for more free-form 1:1s that in-person allows. The latter are key to building trust.

I'm finding this to be one of the key drawbacks of remote work - some decisions require cross team trust and I've yet to see anything besides more frequent ad-hoc conversations that builds it. The problem is that the most trust building conversations tend to be "off-topic" conversations that just don't happen organically in a remote environment. I as an individual can try and make these happen but I can't force the org to do so.

If anyone feels like they have a solution I'd love to hear it (we're already doing quarterly off-sites, it's not enough)

> it suggests there is something about that person that does not align with that.

I think this is insufficiently nuanced (assuming you wouldn't reject someone for going on a lot of dates and being picky). Many people have come to the conclusion that sexual chemistry is really important to a good relationship and thus try to test that as early as possible. This is quite different from sleeping around "for fun".

I've only ever been interested in long-term committed relationships and racked up a high-ish body count this way. One to two dates, clear it's not going to work out, keep looking. I've now been in a long term relationship for over a decade and imagine I'll be in it till I die, many of my friends are in similar positions.

I think we have similar priorities, I just find github to be an extremely poor predictor of on-the-job communication.

I find the signal from this is still too low for it to be worth looking into. Quality of documentation has a lot to do with:

1. Correctness

2. Matching the content to the intended audience

I find that in general I have no time to check correctness, nor is it obvious who the intended audience is. A blurb and basic setup instructions is fine if the intended audience is the author. A tool intended for a particular set of power users may not require motivation. The author may well know how to write more accessible documentation and chose not to for a sensible reason.

As someone on the hiring side, Github projects only matter when they show impact via serious industry adoption (this is not sufficient, leftPad won't get you hired).

The problem is that a Github repo is too much time an effort to evaluate on technical merits. And even if the work is good, how fast can you do it? Some people turn out good work at 1/3 the rate their peers, I don't want to hire them. Between other parts of the evaluation I can give the repo 30-60 minutes max, just not enough time.

> Other states pay for addicts/homeless to be shipped off to California

This has happened but it's not substantial, most homeless in CA were living here and homed before becoming homeless[0]

> San Francisco has a drug problem simply because it’s one of few places in the country where it’s safe to have a drug problem

I've spent a lot of time volunteering with the homeless in SF. This doesn't match the data or anecdotal experience. Many drug addicts follow the path of becoming homeless (overwhelmingly because of cost of living) and transitioning into hard drugs to cope with the pain of living on the street.

[0] https://homelessness.ucsf.edu/our-impact/our-studies/califor...

Uber as a company started making a lot less sense when it became clear they wouldn't crack self driving cars (remember when they did self driving?). The current moat is far less defensible.

I've worked on the CS side of new material discovery.

The tricky thing is that you don't actually have that much data so ML is not even close to plug and play. If you want to get results you end up needing to pair ML with a lot of theory and some tricky algorithms to help narrow the search space and even then that space is huge.

Progress is being made but I think we're still at least 5-10 years from CS providing a real inflection in materials discovery.

My proposal: More or less what we had pre 9/11. We solved the the vulnerability that 9/11 exposed in under a year - the inability to lock down the cabin to protect it from hostiles. Every commercial plane has newer doors installed now.

This means the only really viable attack on aircraft is to blow them up mid-air which:

1. Is much less impactful

2. Still achievable by a determined actor today

It shouldn't because the TSA does not serve a real purpose other than to be a jobs program.

There has been a great deal of evidence over the years that the safety measures of the TSA are easily bypassed (google "TSA security theater" for a wealth of sources) and there's no evidence of increased safety since the department was created.

This is such a weird disconnect between middle class white collar workers and the rest of the country. I know a bunch of people working 60-80 hours a week, usually doing multiple jobs, just to make ends meet. It feels strange to be offered the chance to work less than half as much as another human and receive substantially greater rewards for my efforts. It would be interesting to see what this does to social cohesion.

It can be, but doesn't have to be. "High performance" is just too vague to try and draw inferences about individual companies.

I work somewhere with what I would consider a high performance culture. We all show up to work and consistently put in ~8 hours of high effort work. A few times a year that may mean some late nights, but mostly not, and folks usually take 4-8 weeks of vacation a year. Working around the clock is not rewarded. I find the whole setup fun and invigorating.

That said we do sometimes make a bad hiring decision and those people may end up working 80 hours a week to try to match the output everyone else. We have to let those people go.

> Usually, the problem is that companies find out that I'm not good as a cog in a machine

IMO this is the real problem - and it comes through in your blog posts. It's very risky to hire someone who can't be a cog, you need a lot of social proof that they're worth the risk.

It's funny, the Ivies are actually quite up front that they're goal is to pick "future leaders" of the country, not the most academically gifted students (though this is an important sub-group). The fact that there's a quota of students who are chosen for purely academic talents is particularly telling.

If you have enough drive to get good grades and your parents are worth many millions of dollars you have an excellent chance of being one of society's future leaders, you're they're perfect candidate.

AI doesn't have to do everything, or be the best at everything, it just has to do some things well enough, or augment humans enough that fewer people can do the job.

There's a lot of labor destroying potential in animation. For example, a human could make a small number of key frames (this already happens) and AI could fill in the in-between frames (this is where a lot of labor goes) you could cut costs significantly.

I'd hoped the pre-requisite that titles matched duties was implicitly understood in the message, sorry!

But you can absolutely come into an engineering role at a company with 40 engineers (40 person org charts to director at BigTech), grow into leadership of a 200+ person engineering (charts to vp) and then go back to BigTech for bigger bucks.

> Referring to the "small startup experience" not mattering - I would say it is almost like an alternate track as an engineer.

It can be but doesn't have to be. A lot of people prefer to hire folks with some small startup experience because such people almost always better at thinking about businesses holistically. Many people who spend their entire careers at big companies can only reason about their little slice (lots of bad takes on businesses here).

A common way to make VP of engineering at big tech is the following 5 job hops:

1. Manager at big tech

2. Manager at startup -> promotion to director

3. Director at big tech

4. Director at startup -> promotion to VP

5. VP at big tech

The problem comes when someone gets too senior with relevant experience at scale. I (generally) can't hire someone with 15 years of experience who's never worked at scale. They're going to be bored out of their mind with mid-level work but simultaneously I can't actually trust them to not screw up something in senior+ roles, either technically or socially.

IME in cases where a PIP is truly triggered by performance, 9 times out of 10 the manager has already tried very hard to address the issues and not seen results. It's unusual there's anything to be done.

I will see that as someone who mentors a number of engineers, some of the PIPs I see are ridiculous. Here's a verbatim quote from the PIP of an engineer at a decacorn: "Employee must not receive more than 3 pieces of corrective feedback on any single PR". That's _basically_ constructive dismissal.

> There is a little nuance with respect to setting up a project idiomatically, but once that groundwork is laid anyone can come along as they're simply going to copy the structure and style of what is already there anyway.

You might think so! Having seen this tried in multiple 100+ engineering orgs, I think empirical evidence is against you. CLI tools can get surprisingly complicated (or use a surprisingly complicated set of language features). Many engineers will simply give up if a solution is not obvious in the first 5 minutes.

The end result is you have a lot of team A asking team B for a minor feature or fix to team B's CLI tool, or worse, team A just writes a hacky workaround to the broken-ness in the CLI tool.

Most people just don't care enough to put in the kind of effort for anything but the easy path.

> because it happened once (to Pinball)

Honestly I don't think pinball is a similar example, light reading indicates the market for pinball was in the 10s of millions. Inflation adjusted it sounds like the pinball industry was 1% of what modern video games are and, IMO, that kind of difference in size is a difference in kind.

A better analogy might actually be liquor and prohibition, but we're not seeing anger anywhere near that level

> politicians are just as likely to use a sledgehammer to the entire industry than a scalpel to the worst offenders just like the Pinball Prohibition.

This is unlikely simply due to the size of the market. Video games are played by >60% of the population and generate more revenue than movies with hundreds of thousands of jobs at stake.

Nah, different supply and demand for the two groups.

After the layoffs a glut of senior management, with more directors than you can shake a stick at. Many have had to take jobs as line managers. Director is a pretty similar level to a Senior Staff engineer, the latter is still in relatively short supply.

My question was about promotions - other engineering professions definitely have promotions, and those promotions come with >3% pay rises.

Also, even without boom times top engineers will still make a lot of money because top positions are still insanely hard to fill. I know of several top companies with L8+ positions that have been sitting vacant for over a year, competition for these candidates is fierce.

This is a good analysis, but you're missing another angle: quite a few kids are not developmentally ready[0] for algebra in 8th grade but most are ready in 9th grade. By forcing kids who are not developmentally ready into algebra you generally do more harm then good (most of those kids will hate math for the rest of their lives). This problem also occurs earlier in school life with fractions. When I was in education I saw pushing students too early like this do a lot of damage that had to be cleaned up by tutors at a later date and was totally unnecessary.

I even think Calculus could still be achieved by the end of high school with the right curriculum choices - you can teach a lot of calculus concepts to grace schoolers if you elide the actual mechanics of computation. Anecdotally, students who were pushed through that sort of curriculum earlier in middle school tended to breeze through high school calc.

[0] https://en.wikipedia.org/wiki/Piaget%27s_theory_of_cognitive...