thot_experiment

I agree with the sentiment, but while New Glenn is likely a more economical rocket, it doesn't even have the same payload capacity that Falcon Heavy had eight years ago. They are nowhere near a meaningful competitor yet, and if Starship actually gets up and running the mass to LEO gap between the two systems is like 5x+ with Starship aiming for full reusability while NG expends the second stage. Ofc even whimpy competition is still competition, and BO does have deep pockets.

Crazy that such a load bearing job isn't better funded and more respected. Arguably the most important job in society and the level of respect, pay and to some extent training (at least a lot of places require a masters for what that's worth) is absolutely not commensurate with it's importance.

I dropped out of high school for the same reason, I had a teacher that failed me for writing an essay in three different styles of handwriting, and it just broke me. I wasn't a particularly good student, and I especially had a habit of just not doing essays, but I was making an effort to make it through the humanities and get my shit together, and to have that effort rewarded with a 0/100 just made me view the entire system as an absolute joke. I have a more nuanced take now, but it's still impossible to wrap my head around how comfortable people are with the education system here.

Society is made of people, people! You live in a society. Why do we not want the foundational atoms of it to be the best they can be? It just seems so obvious and simple and non controversial.

Who else has that Tux plushie tho? I've had one since I was like 11 years old.

Sure, I definitely agree that the highest impact work would be to shine the light on the corruption among the leadership and hold them to account, in all cases. However in the institution of American police the corruption is endemic.

Just sharing what I know to be an extreme minority opinion but for my money, cutting up SF traffic on a bike is the greatest feeling in the world. On some level I have a deep respect for the cars because they provide me with a matched partner for the dance. That being said, yeah, ban cars, fuck cars.

I spend 10-20 hours/wk on the bike and I personally hate almost all bike infrastructure. I understand why it exists, and I'm pro it being built but personally miss me with that shit. I hate the separated bike lane more than anything, even practicing parallel bunny hops every ride it still sketches me the fuck out to be trapped in a little channel like that. I'd much rather swim in traffic, and have all the options that that gives me.

I don't know if anarchy helps in this situation, I actually think you need robust social systems with buy in from citizens to prevent the natural accumulation of power. The fundamental problem is that there's a diminishing cost to acquiring power as you acquire power, this relationship should be inverted. The more powerful you are the harder it should be to get more powerful.

This is basic engineering, you don't want runaway feedback loops, the underlying system is unstable so we need a control system.

It's very easy to get started on this, you tax the shit out of people who have a lot of money because the old adage is true.

You're so close! Instead of patching the issue maybe let's solve the root problem of spiky power distribution among humans. We don't need to make sure cops have immunity to prosecute powerful people. We need to not have powerful people.

(though realistically speaking yes there's probably some level of procedural immunity that probably makes sense, similarly with business bankruptcies not ruining the people who start the business)

It's absolutely not the slightest bit crazy if you've paid attention to how cops behave at any point in the last history of the country. 100% agree about personal responsibility. You must understand that when the cops says that oversight means they can't do their job, that means they view their job as bullying, harassing and killing citizens, so yea, we should put a stop to that. 1312

I assume they need to do a bunch of WebAPI bullshit to get around Youtube's draconian policies, but maybe one day https://txikijs.org/ will solve all problems with embedding javascript. I believe, and maybe the strength of my belief will be enough.

the kids are alright

I've wanted a reliable tool like this for the longest time because I'd like to be able to draw arbitrary poses off my phone if I'm just sketching on the train and I'm having trouble visualizing how a shadow would fall or something. I of course would never pay a subscription for it, that's absurdist comedy. What I want is github pages page you can ctrl+s and keep forever, might just put my money where my mouth is on that one too. I've got a lot of three.js under my belt and local models have made me unreasonably bold.

idk, my 10 year old makerbot 2 has been pretty reliable, ever since Prusa slicer came out and I tuned a profile for it maybe 6 years ago it's been spitting out quick dimensionally accurate prints. i use it all the time, probably go through a spool every month or two and all i've had to replace is the cooling fan for the extruder once

i'm mostly printing small mechanical parts and i can't say i have any complaints, i assume a modern prusa would be much better, surely there are other FDM printers that are good?

I can't tell if this is sarcastic or not, but it seems insane to let the AI write the tests.

AI can't be held accountable, it shouldn't be writing the tests that determine whether car systems function correctly.

Yo, MTP for Qwen is sick, thank you! Your work is invaluable.

Also you can feed it ALL of your data willy nilly without ever worrying about safety because you can just do it with the LAN cable unplugged, for applications that demand data hygiene it's a cheat code that guarantees safety without any sort of data sanitization.

Yeah, thanks, though I think local models are at least a Cessna, which while being nothing like an F-35 can fly.

Maybe a skill issue but they both feel about the same and the MoE is 3x faster so I barely use the dense model.

This is probably a precision thing, I think there's a really big difference in long running tasks between q4 and q6.

Sorry, "essentially useless in the context of local model availability". It's a fine model but it's tier of inference is fully fungible.

It's more complex than that, I think the reality is that there's a lot of code that's just not that deep bro. I have some purely personal projects that have components that I don't understand anymore, I wrote that shit by hand, they still work but I haven't touched that shit in years. There's a lot of code that AI can write that's like that that helps me, the stuff I would forget about even if I wrote it by hand. I think you have to have discipline in it's use, it's a tool like any other.

AI, and especially agentic AI can make you lose situational awareness over a codebase and when you're doing deep work that SUUUUCKS, but it's not useless, you just have to play to it's strengths. Though my favorite hill to die on is telling people not to underestimate it's value as autocomplete. Turns out 40 gigabytes of autocomplete makes for a fucking amazing autocomplete. Try it with llama.vim + qwen coder 30b, it feels like the editor is reading your mind sometimes and the latency is so low.

Overall using screentime as the metric, derived from some imperfect logging and vibes it's about 50% OpenCode 15% Continue 15% my homebrew bullshit 13% Claude Code and 7% Cline. I've been deep on agentic stuff lately (1.3wks aka 3 months of AI time), there are only so many hours in the day to duplicate work and AB test, but in the past I've sworn by Qwen Coder + llama.vim and I still enjoy that workflow for deep work far more than I like prompting agents, but there's a lot of dross I'm learning to delegate.

No, it isn't. I am saying that the set of tasks that can be completed by Opus 4.7 has a surprisingly large overlap with the set of tasks that can be completed by Gemma 31B. It is meaningfully equivalent in many cases.

(of course if i'm being honest 640kB is fine, i'm sure tons of the world's commerce is handled by less for example, the delta between a system with 640kb of ram and a modern one is near nil for many people, the UX on a PoS terminal does not require more than that for example, the hacker news UX could also be roughly the same)

Benchmarks only give you the roughest idea of how models compare in real world use. They're essentially useless beyond maybe classifying models into a few buckets. The only way you gain an understanding of something as complex as how an LLM integrates with your workflow is by doing it and measuring across many trials. I've been running Opus 4.7 in Claude Code and Gemma 4 31b in parallel on projects for hours a day this past week, Opus 4.7 is definitely better, but for many things they are roughly equivalent, there are some things on the edge that are just up to chance, and either model may stumble across the solution, and there are some areas of my work that reliably trip up both models and I get better mileage out of writing code the old fashioned way. I understand that I'm just one data point, but I'm not writing CRUD apps here, I'm doing DSPs and weird color math in shaders, I don't think any of it is hard, and the stuff that I think is hard none of the models are good at yet, but idk, they just don't seem that extremely disparate from one another.

FWIW I think Gemma 4 31b is more likely to be of use to me than Sonnet, idfk, maybe it's a skill issue but I love Opus 4.7, undisputed king, but Sonnet seems borderline useless and I basically think of it as on the same level as Qwen 35b MoE.

I 100% agree with your philosophy but I wanna note that I genuinely find Gemma 4 31b to be better than Sonnet. To be clear, this makes NO sense to me, so I'm probably just high and making stuff up or just biased by a small sample size since I don't use Sonnet that often. I find that Gemma 4 makes the sort of "dumb AI" mistakes Sonnet makes less often, especially in agentic mode. I genuinely don't know how that can be true but Sonnet feels much more like "autocomplete" and Gemma 4 feels like "some facsimile of thought".

False. The absolute capability is irrelevant, with the proper harness 31b is more than adequate for a very large portion of the tasks I ask AI to do. The metric isn't how good the model is at Erdos Problems, it's how reliably it can remove drudgery in my life. It just autonomously reverse engineered a bluetooth protocol with minimal intervention, it's ability to react to data and ground itself is constantly impressive to me. I do a ton of testing with these models, today I had Gemma answer a physics problem that Opus 4.7 gave up on. With a decent harness and context the set of tasks where their capabilities are both good enough is very surprising. The tasks I have that stump Gemma often also stump Opus 4.7.

No, exactly the opposite actually. Qwen3.6 is too imprecise for long running agentic tasks. It doesn't have the same ability to check itself as Gemma does in my testing. I keep Qwen MoE in vram by default because there are tons of tasks i trust it to oneshot and it's 90tok/sec is unparalleled, anything where I don't want to have to intervene too much it can't be trusted.

Flat wrong. Q6 Gemma 31b feels a lot like opus 4.5 to me when run in a harness so it can retrieve information and ground itself. The gap is not that big for a lot of usecases. Qwen MoE is fast as fuck locally for things that are oneshottable. I have subscriptions to all the major providers right now and since Gemma 4 and Qwen 3.6 came out I haven't hit limits a single time. I'm actually super surprised by the number of things I try with Gemma 4 with the intent of seeing how it fails and then having Claude do it only to come away with something perfectly usable from the local model.

Depending on your laptop, if your laptop is a Strix Halo or a Macbook with a decent amount of ram, that day they arrived is about 6 months ago, and today if you can run Gemma 31b, you're golden for your basic workslop code. You can do most of it with local models. Heck, for a lot of the tier of programming you might encounter in the average job Qwen 35b MoE is good enough and it can hit 100tok/s on decent hardware.