Viewing profile — ThunderSizzle

This must be the new linked in strat. What I learned about using the best model after talking to a homeless person.

So you buy their sugary crap to hide the shit coffee. Pumpkin spice lattes.

That's me. I've just moved the puzzle to figuring out the CC optimal "language" and mode of working with an LLM. At the same time, I haven't lost my love/expertise... Yet. I explai…

Well, thanks to evergreen and auto updating systems, there's no guarantee that X app will be the same today and tomorrow. It's probably what I hate most about Android- apps shouldn…

I have issues with the vulkan backend. It seems to cause over all GUI sluggishness, but maybe that was actually Firefox. I'll give it another go.

An AMD R9700 gets 20-50 TPS at ~300 watts on 27B. 100 TPS for the 35B MOE model. And there might be some more optimizations to that as AMD software support gets better with ROCm's …

Is two generations ago before 2020? Everything is 100% more expensive than pre-2020 prices. Eating out is there. Power Tools are there. Land is beyond that and Housing has been the…

I've never purchased rental insurance from a landlord or their recommended place. I used my insurance company I already had.

I think the point is that an i7, even from ~15 years ago, is probably sufficient for most gaming needs, especially if the game is primarily GPU-bound. As long as you give the older…

Meh - our dev machines are over-provisioned cloud pc's from MS, that run as fast a laptop from 2005. And I'm sure they are paying a premium for the privledge. My 10 year old deskto…

Massive immigration would cause the system to completely collapse as imported people will not respect the local economy, culture, etc. in the same as natural citizens that original…

If I'm picking a stock to buy (in the "retail" market, it's primarily based on a balance of EPS, P/E ratio, and a low(er) amount of debt. My P/E filter filters out the likes of Nvi…

The obvious brain-child of this concept is HP and their consumer inkjet printers. Originally, they sold their printers for "cheap" and then sold ink for a premium. However, as cons…

Which is probably a backwards anti-pattern companies have built. Your most expensive engineer's time is most valuable, so if you give them standard issue which is half the speed, y…

Luckily we're not, because then your tip to a waiter/waitress would be dependent on their student loans remaining, especially considering how many expensive liberal arts majors str…

Catastrophies would be we vibe coded a nuclear plant or space rocket system and we blew up thousands of people due to a vibe coding error.

I see this said a lot, but Amazon products I buy actually work (even if the quality of obviously laughable - it actually performs the function and won't break any time soon to just…

They'll do local LLMs you have to pay for. Best of both worlds: your LLM processing power will be locked behind a subscription. You'll be coerced into a subscription to unlock the …

I'll give 27B-MTP a try. I think I can tolerate 45 tps if the results are technically better. 35B is pretty good, but definitely shows it's inabilities at times (probably either du…

Agreed. I have a single 9700 and I'm able to fit Q6 27B at 30tps or Q5 35B at 100tps very easily via llamacpp running vulkan. The results are impressive considering the amount of p…

The cheapest 3090s I could find with any sort of guarantee were pushing $1500. An AMD AI Pro R9700 32GB brand new is $1350 right now. After some tweaking, I had it running faster t…

Obamacare, but for AI, where every American has to now pay a penalty to not use AI or something like that? That was the last major thing the Democrats did, and healthcare has gotte…

The CEO won't get in trouble, but the employee who can't justify a bad result/prompt?

There will always be a competitor that can undercut the inference market. There is no "moat" given that you can self host decently capable LLM agents like Qwen3.6 on not super expe…

An R9700 is $1350 and can get 100 TPS running Qwen3.6-35B-A3B Q5 with 130k context window (with room to spare) with a bit of fine tuning llamacpp-vulkan, but llamacpp's repository …