thot_experiment

In comparison to some absurdist baseline maybe, actual software NEVER stops working under you, so in comparison something like an works 80% "most of the time" is godawful. Though I would argue that with SaaS the trend is toward 100% likelyhood to fuck your shit up given enough time, and it has borne out this way in the real world time and time again. SaaS is popular because it allows companies to more effectively extort you for your dollars.

It's a SaaS, when in the history of SaaS has it ever been a good idea to trust that the company won't ruin the product under you?

Ahhh, I see the issue, you must be using a Windows other than Windows 10 LTSB and presumably you don't trust your router. I dual booted for a long time before VMs were good, but the user experience of booting over is just so awful.

If you mount the Linux OS volume under windows while it's running it can get all fucked up because the two operating systems are not chill about who's really in control. You can avoid this by keeping the linux OS on it's own volume, but putting all the user data on a separate volume, you CAN mount this under both operating systems and get good speeds under both OSs, wheras using WSL native ability to access the host NTFS system is very very slow.

Dual booting is hell, it's never good, it's never been good. I don't use my computer in a modal way. Willing to accept booting over as a friction point is crazy. You can just disable windows update and all the nags and then it's just a good operating system, mine often runs for many months without reboot. I have never seen a single ad in my OS.

It's absolutely not the best of both worlds if I can't simultaneously game (or run CAD software in my case) while running torch w/ triton at the same time.

Dang, given all the cool visual work on your website I absolutely would not have expected that. Always good to be reminded how disparate we all are from one another.

Makes me grateful I got my mom to sign that waiver to let me get on Neopets, I don't even see the hex code anymore, I just see marigold, umber, vermilion.

I hope they're aware of this:

https://www.nature.com/articles/d41586-026-00945-7

You can't clone forever.

You're so close to getting it!

The best option is to run a Windows host with WSL and an extra EXT4 volume mounted to both windows and linux (this is a non issue as long as it's not the boot drive for linux). User WSL for everything linux, optionally you can also just mount the EXT4 volume with all your stuff to a real VM if you need to (this lets you have windows GUI apps in windows, tho maybe WSL does this now too?). Everything works, you have the best of all worlds.

Do you only play singleplayer games?

Nice, FWIW this is currently pretty easy to solve by just keeping your stuff on a separate EXT4 volume and then mounting that under Windows. Windows accessing a mounted EXT4 volume through WSL is much faster than the other way around.

Because it's like 5x cheaper and beats it out spec wise in almost every way?

Totally understandable. YMMV but I found the llama.cpp build process to work on the first try on my machine, and it only takes a couple minutes, which definitely isn't my usual expectation or experience. I was very pleasantly surprised. Their web-ui is also getting very polished while still doing a great job of letting you tweak all the weird settings.

I think comparing companies to Apple is kinda silly as they're unique. They're in a position of such market dominance that they can afford institutional arrogance.

However, it's very illustrative in this situation because I think GoPro made the mistake of thinking they were Apple, and at their outset they were compared to whatever other dross was on the market. Unfortunately it went to their head (almost certainly they were run by MBAs instead of product people or engineers, or anyone who would actually use an action camera seriously) and they completely lost sight of what their product even was.

I think it's completely avoidable, they could have just kept trying to make good products instead of whatever subscription service bullshit they attempted. They saw they had market dominance and so they tried to rape their customers as best they could because they didn't see the customers as having a choice. Don't do that. They cashed in on their market position for short term gains and it destroyed their future.

This is so sick except it only has 2 pulse counters instead of the 4 on the S3 which means I can't use it as a drop in replacement on my current project. Not really complaining, I cut my teeth as an embedded dev on the ESP8266 and for years now all of my personal projects (and a fair few professional ones) have been based on the ESP32 line of chips. They're all pretty incredible for the cost, absolutely my favorite embedded target.

I just build llama.cpp from scratch on the PR that has MTP drafters.

https://github.com/ggml-org/llama.cpp/pull/23398

Please don't use Ollama, it's a bad actor in the OSS community.

They certainly wouldn't survive if capitalism was working as intended. GoPro is one of the most institutionally arrogant companies around, they seem unable to understand that they have competition. (Which to be fair they absolutely didn't for the first few years of their existence). In 2026 they make a mediocre product with horrible software, and even worse customer service.

In contrast, I've smashed the absolute shit out of my insta360 lenses 3 times now (total cost to replace lens 3x $50), and when I smashed the screen to smithereens, they fixed it, OUT OF WARRANTY, for $50 including parts and labor, and they paid for shipping.

There is zero reason to buy a gopro, they don't have a single compelling product and HOLY FUCK the software is bad compared to insta360.

You can get SXM V100s for like $100 off ebay, if you're willing to do the troubleshooting work to get em running with adapters you can build a computer capable of fitting a Q4 quant of a 120b model in VRAM for something like fifteen hundred dollars. (assuming you already have some RAM sticks laying around T___T)

Ollama is a shitty project that steals from the open source community, don't use it, use llama.cpp instead.

Hard disagree, Qwen multimodal is way better than google's, but Gemma 31b runs laps around Qwen 27B in complex engineering tasks. Maybe Qwen is better at slopcoding web framework CRUD, but for embedded dev there's no comparison.

I don't know about this model, but the next one up, the 31B I've been using as an agentic coding assistant in OpenCode, and basically anything that's easy enough that I'd trust Sonnet to handle, I trust Gemma 4 to handle and it's been doing a great job, it surprises me positively much more often than negatively. I not infrequently run into situations where Gemma 4 fails to do the task and I switch to Opus 4.7 and it fails also.

I've always found the Gemma models to vastly under-perform on vision tasks compared to Qwen so that's nothing new.

I haven't tried this model yet, but I can run Gemma 31B w/ the MTP drafter in pure CPU at about 10tok/s so this should run at about 20-30tok/s on a decent CPU, it'll probably run at >50tok/s on any Mac that can fit it, and lots of people have a gaming GPU with enough VRAM. In terms of access to hardware being a gate, it's one you can hop pretty easily.

I think resiliency to the corruption isn't evenly distributed among the populace, there are definitely those people who are capable of resisting it long enough to get shit done. I just think it's mostly the levels of abstraction that these people have between them and anything real.

Honestly, if you suggest surveillance state shit, you should just be mandated to livestream your life 24/7, you have nothing to hide right?

It's so frustrating because even the notionally good actors look at this and think "yeah but it's not gonna corrupt me"

No you fuck, the only way to walk the good path is to give up power and never let it concentrate again.

Larry Ellison of course is one of the most anti human people on the planet and I doubt he was ever anything else.

Love seeing AMD forcing Novideo to catch up for once rather than the other way around.

lmao, you're completely out to lunch, my friend is a teacher in one of the most well funded districts in the entire country and it's a decent job but it's an incredible amount of work and he's not making an amazing wage considering how much he needs to work outside of school hours

i can only imagine the horrors faced by teachers even on the other side of the bay, much less in a red state that isn't the fourth largest economy in the world

tenure protections can be problematic, but so are activist parents, the system isn't great but it is necessary to some extent, your exposure to bad actors as a teacher is massive and requires commensurate protection, just stochastically you'll get parents trying to get you fired every once in a while

q6 is fine for that qwen with ctx @ q8, and the dense models of that size are solid at q4 with q8 ctx

Naw totally agree, we need way more robust protections for security researchers and way harsher penalties for corpos doing bullshit, it should be a percentage of revenue.

We have way too much fuck around these days and not nearly enough find out.

Tried this extensively in OpenCode, never used it once since Gemma 4 came out, got into thought loops and did stupid edits I didn't ask for more often than the local 31b model. One of the worst "frontier" models I've ever tried.