Once people realise they solved the hardware limitations of AI models with software, it's game over for AI stackies. They have no moat either, AI made Ai companies obsolete.
Hmm so when can I achieve similar results while having 64 or 32gb ram
>>62930095With 32GB of System ram you can get away with Qwen3.6-35B-A3B if you have a 8gb of Vram gpu.I have a bit more horsepower and running Flash-next at Q3 which is good.
>>62930072>freeIn what way are you the product?
>>62930154Might be free for now to catch new fish.
>>62930154>MIT License >Open codeNo catch, Just one based lad, that is all I'm going to say.The guy did it for Tech Rizz,
why no one has created a share pool of gpu of people wants to rent on the internet to use opensource models yet?
>>62930227You mean literally every single cloud computing provider?
>>62930154When you updoot the github page it looks better on the devs' CVs, kek.
>>62930135I have a business laptop with no Video Cardano. I stopped using my PC about a year ago because there's no reason to use gpu. (Modern games are bad)
>>62930247yeah but using people desktop, install, set a desired price(price will be derermined by region), let it run, check balance., thats it.
>>62930154We've been the product for almost 30 years by putting all our collective thoughts for free online.
Thing is I dont trust these local Queens I am a gemini maxxer gemini is My Favorite AI I cant stand chippity and clød
>>62930227I understand what you mean, the problem is most consumer PC's are not good enough for the type of work however a Strata like hybrid cloud system would unique idea, would make all the big cloud providers seethe! haha >>62930278>Modern games are bad100% with you on that lad, but unless you have newish cpu with good iGPU you'll struggle. Best bet if you want to dip your toe in the AI space is put $20 into openrouter, download opencode and mess around with some cheap chink models. But you wont be able to do any based questions.
>>62930072Iexec already solved in using decentralized cloud computer. Digital oil chads won
>>62930154>anon learns about open source for the first time
>>62930299>Dont trust uncensored open sourced modes>Uses JewAi made by (((google))) Okay bud, you the type of guy to use the tampons in the Mens bathroom? hahahaha
>>62930432I dont trust the answers to be correct. I have had best results with jewgle's AI. I dont care what they know about me because I cannot be blackmailed as NEET
>>62930489shhhhhhhhhhhh! ;D
>>62930411Dont know anything about that anon, tell me more.
Now this is based, glad I didn’t splurge on an m5 studio/max whatever the fuck they’re called. I guess I should upgrade my ram to 64 at least (REEEEEEE). t. 7800x3d 4080 super and 32gb ram ddr5 cl30 nigger here
>>62930530The founder is a French psychopath who loves tranny hookers
The qwen distills are all too stupid in the name of being fast on potato hardware. Even if you only use them for straight coding and they are exceptionally stupid if you ask them to branch out to anything besides that.
>>62930072I really think people don't grasp how insane this is. Its like having your own in-house supergenius. But not just that. Your own unlimited in house supergeniuses.
>>62930135Qwen 4 27b is going to be released soon
>>62930558Dont upgrade, dont give them your money anon. Just use Strata Qwen3.6-35B-A3B, it would run very well on your set-up. >>62930594?!?!?!? I've never used a base Qwen model so I never know, I use uncensored models which always answer correctly. >>62930605keep back-ups anon, once "they" realise how based this is they going to make it... you know what I mean. >>62930667The main Chink left a while ago, dont think it's going that much different anon. It be better if they released a new 35B-A3B but maybe a 38B-A6B active version.
>>62930707I can ask qwen>go look at public 15 minute delay open interest on this expiry>calculate naive gex on %TICKER>display naive gex in a chartAnd it will shit itself 100 times out of 100. Gemma is a little better but only because it's a little better at initial data gathering. But anything /biz/ related that's more complicated than 'you can just eyeball it so why ask a model anyway?' it's too stupid to go through a few steps regardless of context size.
>>62930072Just wait for "Iranians" to take down the power grid using "open models"Ve Muzt Ban Open Models IMMEDIATELY
>>62930154Its just software, you still have to have compute power
>>62930707I’ll give it a shot anon, I need it to be capable of being (((antisemitic))) as well as parsing through my vast collection of pdfs/research studies on Rockefeller medicine to help me take down the system
>>62930393Kek did you read the answear, the military will protect the governement and flood the country with more brown you absolute retard
>>62930072they really don't have a moathttps://x.com/realsteveeisman/status/2122766726265700774
>>62930072Whats the difference between this and claude code or codex app running with some skills on local qwen 3.8
>>62930393i mean, theres lot of crypto miners that would pivot to this if they need
you guys all have 24GB VRAM GPU's sitting around to run this on? That's still very much inaccessible to the average consumer.
>>62931064Strata is inference engine. You can use it with Codex and others.
>>62930227What do you think bitcoin mining is..