DirtyWaterHotDog
No bio...
User ID: 625
DirtyWaterHotDog
1d ago
Yes ! Probably the fastest I've seen a set of tools be adopted. It's the gold standard now.
In my new job and old job, we used both uv and ruff. The move to uv took a bit longer in the new job because it involved changes across 1000+ engineers. But it got done. Ruff integration in both cases was trivial.
uv was transformational. It is a great tool, yes. But a big part of it had to do with the dire state of python packaging it replaced. Another part of it had to do with the drop-in nature of it. The porting experience gave me a ton of joy.
ruff is great. It primarily solves annoyances. Some people still use flake8, black, isort and 20 other tools. But most greenfield projects are starting with ruff. But now that ruff is popular, you can share and steal complex linting/formatting logic in public to make it more powerful.
ty is new. Technically still in beta. We use based-pyright which is also new. It's stable and works. But we only run based-pyright as a pre-commit hook. ty is 10x faster, so once it is stable, we will be able to run it aggressively on saves. We've tried ty internally and senior devx people are excited. But, we're waiting for it to reach 1.x major version before making the port. Majority of python repos either don't have type checking or use mypy which is about 50x slower and annoying to use. So most team should see a bigger improvement than what we'd experience.
If I to guess, astral wants to work their way up to a JIT compiler for python (like pypy). If the linter and type-checker can enforce strong code behaviors then a JIT compiler should technically be build-able for python. But, the future is any one's guess.
Python is going through a devx revolution right now. Pydantic, Astral and Mojo are the main contributors.
Mojo is typed, compiled and a (claimed) super set of Python. It hasn't seen as much adoption, but has is led by systems Jesus - Chris Lattner. I'm hopeful it will get there eventually.
Astral on the other hand, has transformed the python dev workflow. 'uv' solved python packaging. 'ruff' solved linting and formatting and now 'ty' solves python type-checking. Separately, Pydantic allows data objects to be strictly typed and is pretty much a python built in.
And I know it's customary to throw a bunch of half-baked tools at someone to silence criticism about a language. For years, that was true for python. But no, these tools have genuinely become ubiquitous. The python code-base at my current job is pretty much strictly typed.
In a few years, I'm betting python will become a pleasant language to use.
DirtyWaterHotDog
1d ago
viable option for anything that even vaguely cares about performance
Bit Hyperbolic no ?
I'd say the opposite. GC languages are only unviable for systems that care about exceptional performance.
Quant trading works with GCs. ML & gaming have a unique preference for C++ because of the ecosystem, so I'll treat them as exceptions. Google uses Go for large scale systems (not the core, but pretty much everything else). Clearly it's good enough for most systems work.
On DEI-fication of the army :
I'm assuming you served through Bush Jr, Obama, Trump and Biden. When did DEI creep into the military ? Was it a Biden era phenomenon, or did you experience it during Trump 1 and Obama's terms as well ?
The post-Obama picture of America is pretty clear to me. It's the standard populism cascade. One side elects a habitual line stepper and the other party escalates in response. The escalation through Trump 1 --> Biden Autopen --> Trump 2 makes sense.
I see 'Obama --> Trump 1' as a separate phenomenon that's divorced from this escalation ladder. Is that also how Republicans see it ? Or do they see Trump 1 a response to a perceived populist line-stepping by Obama ?
DirtyWaterHotDog
1d ago
Cynical take about the open source programming languages world
The best systems engineers are trans and mentally unwell. Appearing progressive is how you keep them productive instead of spiralling. The 2nd best systems engineers are virgin gooners. Appearing progressive gives them a chance to be around women. The 3rd best systems engineers are m-lady neck beards. Appearing progressive is how they simp.
Everyone else who's good enough to be developing the rust lang is getting paid millions at a quant firm or millions at an llm frontier lab.
Surface level progressivism is win-win stable state for open source PL.
DirtyWaterHotDog
2d ago
Dude, no. I need this.
One day he'll Hank Green this into a blockbuster and I'll be here feeling all hipster.
- Girl at MIT exists
- Student at MIT has girlfriend
- Prof cheats with girl student on MIT boy despite having 3 kids and a wife
- MIT boy with bright future decides to kill professor
- MIT boy acquires gun and kills professor in home (from reports, wife was in the house)
- Whole thing stays secret in dorm
- Involved female keeps silence
Sounds unlikely.
Current guess:
- Family being at home makes break-in unlikely.
- Inside the building removes freak accident unlikely.
- Targeted killing most likely.
edit: gotta be honest, did not have 2 decade old neck beard hate grudge on my bingo card.
DirtyWaterHotDog
3d ago
Some details:
He was killed at 9 Gibbs street, Brookline. I lived within walking distance of this neighborhood. It's exceedingly safe. There are no murders in Brookline. In fact, there was 1 homicide in Brookline in the last 20 years (2006 - Dec 2025), and that was a drug deal gone wrong.
He was killed in the Foyer of his apartment building. Burglars generally prefer single family homes, and and there are many in this neighborhood. Already sus. Next, if it was a break-in, his body would have been inside the house. Why would someone come all the way up to his foyer (inside the building), and then kill him outside the apartment ? Does not sound like a professional criminal.
Pro-Israel, Hanukah and Brown University timing only makes it worse.
Nick Fuentes's ideas have zero intellectual worth. He is a mega-church pastor for the religion of inceldom. Incoherence is key to his movement. At least Candace Owens is schizophrenic. What's Fuentes' excuse ?
Japanese people, Taiwanese, Koreans, or Jews tend to also be pretty great
I'm not convinced that average-IQ is singularly responsible for societal-IQ. They're related, sure. But, IQ differentials have existed for millennia. If the correlation was so direct, then high IQ nations would've achieved insurmountable gaps between them and other nations. This hasn't been the case. Japan, South Korea and Scandinavia are high-IQ regions today. But, they were relatively backwards throughout history. That there is flux implies that IQ is not the primary factor in creating stable and flourishing societies.
Smart Jews make more money than goys
Can't read goy without replaying the meme of a rabbi crashing out at Barney the dinosaur.
you don’t want Blacks/Muslims/Indians in your country
This is anti-HBD. Between Muslims, Blacks and Indians (presumably you mean south asians), you're looking at 4 billion people. Say their average IQ is 90. Let's call them group A. Let's say group B constitutes non-chinese desirable immigrant groups with an IQ of ~105. (I'm assuming most western Europeans do not want to immigrate to the US and Chinese are the rival civilization). Group B will have around 500 million people at best.
Doing some ChatGPT math, Group A has around 90 million people above IQ 120, and group B has about 80 million people above IQ 120. IQ is measurable and group A is more strongly motivated to immigrate. Therefore, if IQ = HBD and HBD = societal destiny, then the US will end up importing a very large number of blacks, south asians and muslims.
Personally, I believe cultural compatibility is just as important as the intelligence of the people you're importing. But, if HBD becomes the primary driver of immigration policy, then it will inevitably sample the cream of the largest (4 billion strong and growing) cohort.
Once you understand HBD, liberals become obviously wrong on most every social issue
No. I increasingly believe that liberal policy is a direct result of deeply internalizing HBD. (By liberal, I mean the American center-left, neo-libs, academic elites and NYT types. Not the communists). You can't tip-toe around landmines this effectively unless you know their precise location. I don't want to derail the discussion, but IMO, American liberals are the result of trying to reconcile protestant ideology with the realities of group IQ.
American Protestantism ties a person's self-worth to their economic productivity. It claims that people are created equal, and given equal opportunity, the hardest working will be the most productive and most moral. Working hard (sanctity of work) reflects good moral character, and the primary observable metric of hard work is economic productivity.
If IQ is real, heritable and puts a ceiling on an individual's productivity, then the whole moral framework stops working. Corporate America stops working. It's impossible to motivate hustlers, aspiring grinders and temporarily embarrassed millionaires if IQ is the primary contributor to outcomes.
If the smarter kid will always do better, then why work harder ? Why put in effort ? If you'll never be able to intellectually compete with the nerds, then why play their game at all ? Why be a peaceful participant of a system that guarantees your loss ? Better to bring it down instead. The natural conclusion is to use populism & violence to reclaim power and set up a tribal society instead.
Liberals need the lower class to believe that they can make it if they try hard enough. So, they set aside a few visible roles for all races, so that everyone keeps believing in protestant morals. They know that in the absence of a 'DEI', the elite will look so different from the base population, that a revolution is inevitable. Liberals want to set up socialist safety nets, because they understand that low-IQ people can't lift themselves up by their bootstraps.
Liberal policy is an uncomfortable compromise of believing in both HBD and protestant values. The resulting cognitive dissonance is why even the smartest liberal suddenly loses 50 IQ points when talking about certain issues.
"It is difficult to get a man to understand something, when his salary identity depends on his not understanding it."
Fair enough. I work with numeric data. Most loops get vectorized as part of the numeric processing packages I'm using. I can imagine there are situations where nested loops can't be avoided on the critical path, and that causes a lot of pain.
More options
Context Copy link