This is a tried and true way of working on puzzles and other hard problems.

I generally have 2-4 important things in flight, so I find myself doing this a lot when I get stuck.

The first time I used Mac OS/X, circa 2004-2005, I was blown away by the design and how they managed to expose the power of the underlying Unix-ish kernel without making it hurt for people who didn't want that experience. My SO couldn't have cared less about Terminal.app, but loved the UI. I also loved the UI and appreciated how they took the time to integrate cli tools with it.

I would say it was a marvelous human experience _for me_.

Sure it was the Apple engineers' and designers' labor transformed into a product, but it was a fucking great product and something that I'm sure those teams were very proud of. The same was true with the the iPod and the iPhone.

I work on niche products, so I've never done something as widely appreciated as those examples, but on the products I've worked on, I can easily say that I really enjoy making things that other people want to use, even if it's just an internal tool. I also enjoy getting paid for my labor. I've found that this is often a win-win situation.

Work doesn't have to be exploitive. Products don't have to exploit their users.

Viewing everything through the lens of production and consumption is like viewing the whole world as a big constraint optimization problem: (1) you end up torturing the meaning of words to fit your preconceived ideas, and (2) by doing so you miss hearing what other people are saying.

The rest are compiler extensions, but if you’re in the space you quickly learn that portability is valued far less than program optimization. Most of the point of your large calculations is the actual results themselves, not the code that got you there. The code needs to be correct and reproducible, but HPC folks (and grant funding agencies) don’t care if your Linux/amd64 program will run, unported, on Windows or on arm64. Or whether you’ve spent time making your kernels work with both rocm and cuda.

And then another index. And at some point you want to ensure uniqueness or some other constraint.

And then you’re rewriting a half-complete and buggy SQLite. So I’ve come around to defaulting to SQLite/PostgresQL unless I have a compelling need otherwise. They’re usually the right long-term choice for my needs.

Maybe on an infinite timescale, but there are plenty of examples of cartels that have been stable for decades or longer.

Hell, even the SV wage collusion was stable long enough to affect plenty of salaries for years.

> There are many resources on how to negotiate. It's worth your while (literally) to study them.

I am not new to this game, and your perfunctory advice of “learn to negotiate” is neither helpful nor relevant to the issue of collusion-through-algorithm.

> Employees are valued according to the value they contribute to the company. Not the profits or losses to the company.

So we agree then the overall financial state of the company is almost completely irrelevant to a salary negotiation. Good.

Which means your equivocation of quarterly filings with past salary information is entirely disingenuous.

They are different, and the company is entering into negotiation with you with asymmetric information. That’s always been the case in a general sense; it’s more so if they know your detailed salary info.

> Companies are not a jobs program, nor a social welfare organization, nor one's parents.

I’m not sure you could be more patronizing if you tried.

We’re all adults here. I don’t think pointing out the disconnect between employer financial status and layoffs / salary negotiation at all implied that I thought anyone was owed a job.

It sure does put perspective on the lie that “a rising tide will raise all ships” though.

You can’t always avoid people who are aggressive towards others, but I’ve found that my life is a lot more stressful when I work with aggressive people, so I actively try to avoid these situations and work in more collaborative environments.

You have CEOs claiming that AI is driving layoffs alongside CEOs of Anthropic and OpenAI talking about the end of white collar work. All this is then amplified by tech journalists like Casey Newton and Kevin Roose. The biggest public proponents of AI keep telling people that it will take their jobs.

What comes after the end of jobs? Who knows. Sam Altman occasionlly making vague statements about curing cancer. There are vague hand-waving notions of a Star Trek utopia.

But to be honest it feels more like a Cyberpunk future, where the Altmans and Musks get to live cancer-free and the rest of us eek out an existence without jobs or any prospect for a better life. Or maybe it looks more like Star Trek, but we're all red shirts.

Can you blame people for hating this?

As we’ve seen time and again this year, highly profitable companies will often lay off workers. So the overall health of a company has little to do with how much it necessarily values employees.

And the tech industry has a record of both using machine learning to skirt laws and a history of illegally colluding to suppress wages. This feels like a greatest hits album for both.

That is my concern: that all of this data combines with some black-box invasive “AI” model and then, magically, all my offers end up being lowball. Not whether I can decline to accept a specific lowball offer.

Actually, bazaar [0] (now breezy [1]) was a distributed version control system written in Python. It gained some non-Python bits over time, but iirc it was originally all Python.

As a (spiritual) successor to Tom Lord’s Arch [2] it was the second DVCS I used and, while slower than git, was performant enough for my needs at the time I used it.

Most distributed version control is IO bound, and Python isn’t terrible at that.

[0]: https://en.wikipedia.org/wiki/GNU_Bazaar

[1]: https://en.wikipedia.org/wiki/Breezy_(software)

[2]: https://en.wikipedia.org/wiki/GNU_arch

This creates an echo chamber where you don’t get reality checks, and when you do they’re easy to brush off as some form of “sour grapes,” after all if the person telling you that you’re wrong was so great they’d have your level of wealth.

I think it takes a really extraordinary person to avoid this. As far as I can tell, most of the modern Silicon Valley titans are not extraordinary in this respect.

I thought it was primarily a user of Anthropic and OpenAI APIs, so the fewer tokens you use to accomplish a task, the higher their margin.

I find myself increasingly nostalgic for the Clinton era. I am not at all sure I will enjoy the version of fuckedcompany that gets vibe coded when this bubble pops.

They’re not pointless; they’re just not the first thing to optimize.

It’s like worrying about cache locality when you have an inherently O(n^2) algorithm and could have a O(n log n) or O(n) one. Fix the biggest problem first.

Once your data layout is good and your cpu isn’t taking a 200 cycle lunch break to chase pointers, then you worry about cycle count and keeping the execution units fed.

That’s when integer tricks can matter. Depending on the micro arch, you may have twice as many execution units that can take integer instructions. And those instructions (outside of division) tend to have lower latency and higher throughput.

And if you’re doing SIMD, your integer SIMD instructions can be 2 or 4x higher throughput than float32 if you can use int16 / int8 data.

So it can very much matter. It’s just usually not the lowest hanging fruit.

Was there a better option? I’m sure. Choices were made. Regrets were had. Switching to uv was a huge improvement for our purposes.

One of the bigger pain points I’ve faced in Python is dependency resolution. conda could take 30-60 minutes in some cases. uv took seconds.

A serious quality of life improvement.

You have to actually care about quality with these power saws or you end up with poorly-fitting cabinets and might even lose a thumb in the process.

If you spend $50 at the arcade you usually develop a little more skill at the game. Depending on the game and player.

$50 at a slot machine develops no skill. At best you’ve broke even or made a little money. At worst, it just feeds an addiction. But there’s no skill here; the odds of any outcome are fixed regardless of what the player does.

Your soldering skill (and sense of adventure) would have to be far better than mine to even consider doing that for wireless earbuds.

It’s probably N.Y.T. style requirements; a lot of style guides (eg: Chicago Manual of Style, Strunk & White, etc) have a standard form for abbreviations and acronyms. A paper like N.Y.T. does too and probably still employs copy editors who ensure that every article conforms to it.

Hard disagree. Borland TurboVision [0] was one of the greatest TUI toolkits of the DOS era, had all of these:

> Turbo Vision applications replicate the look and feel of these IDEs, including edit controls, list boxes, check boxes, radio buttons and menus, all of which have built-in mouse support.

Well, I can’t remember if it had tabs.

[0] https://en.wikipedia.org/wiki/Turbo_Vision