ThunderSizzle

5 ms·

31.
by ThunderSizzle 2mo ago
So you buy their sugary crap to hide the shit coffee. Pumpkin spice lattes.
32.
by ThunderSizzle 2mo ago
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 explained it yesterday that it'
33.
by ThunderSizzle 3mo ago
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't need an update as often as they do.
34.
by ThunderSizzle 3mo ago
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.
35.
by ThunderSizzle 3mo ago
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 latest versions.
36.
by ThunderSizzle 3mo ago
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 there for 6 years now. I'm not surprised. The March of inf
37.
by ThunderSizzle 3mo ago
I've never purchased rental insurance from a landlord or their recommended place. I used my insurance company I already had.
38.
by ThunderSizzle 3mo ago
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 CPU enough RAM, an SSD, and a good GPU, it probably is suff
39.
by ThunderSizzle 3mo ago
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 desktop sitting next to me is a lot faster, but alas, I
40.
by ThunderSizzle 3mo ago
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 originally built the nation. That's not xenophobic, just realit
41.
by ThunderSizzle 3mo ago
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 Nvidia, Amazon, etc, whereas my d
42.
by ThunderSizzle 3mo ago
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 consumers realized any old printer ink works just as w
43.
by ThunderSizzle 3mo ago
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, you are throttling the value you can get from your engin
44.
by ThunderSizzle 3mo ago
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 struggle to find a sufficient career.
45.
by ThunderSizzle 3mo ago
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.
46.
by ThunderSizzle 3mo ago
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 justify a 100x price increase for something more...artisan.
47.
by ThunderSizzle 3mo ago
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 processing power you already have, and it'll
48.
by ThunderSizzle 4mo ago
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 due to the heavy caching quantization I'm doing
49.
by ThunderSizzle 4mo ago
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 people trashing AMD and still trying to recommend 3090s.
50.
by ThunderSizzle 4mo ago
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 than the models the 3090 could run, and it could obviously ru
51.
by ThunderSizzle 4mo ago
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 gotten substantially worse...but at least it's well regulate
52.
by ThunderSizzle 4mo ago
The CEO won't get in trouble, but the employee who can't justify a bad result/prompt?
53.
by ThunderSizzle 4mo ago
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 expensive hardware, like an AMD R9700, and still get c
54.
by ThunderSizzle 4mo ago
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 instability and lack of real versioning frustrates me.
55.
by ThunderSizzle 4mo ago
Like a compliment sandwich or something.
56.
by ThunderSizzle 4mo ago
The coding isn't the hard part. It's the people and networking. Facebook's only moat is HOA boards that think private communication behind Facebook groups somehow equates to public messaging a community... In other words, onc
57.
by ThunderSizzle 4mo ago
I hope this is sarcasm, or "half" my job doesn't exist or something. Or you talking about full time non-dev scrum masters?
58.
by ThunderSizzle 5mo ago
My 980 is currently unpowered. I have not tried to integrate them yet. It's on my to-do list, but the first time they were both powered on, the system had booting issues, and I didn't want to care at the time, since the 980 was pr
59.
by ThunderSizzle 5mo ago
GitHub Copilot. It was one of the best values around in terms of cheap LLM access since each prompt was basically 4 cents (more or less), no matter how much it would do or how many tokens it used. A simple "Proceed" prompt that wa
60.
by ThunderSizzle 5mo ago
I added an R9700 32GB to my 10+ year old desktop that had a 980 4GB card in it, for a grand total of $1350 or so. The payoff compared to what I was using with GHCP was 33 months, but when GHCP announced their price increase, it basically be

More ›