batchsize0 @batchsize0

@landontgreen Is this one of these aliexpress gpu ? If any one has experience with these Chinese good offers are they legit or scam ?

Claude duet has better quality and is more polished. Hermes is less polished works with everything, if you use a local model, privacy but less quality .. Me personally with my single DGX spark prefer local qwen flash instead of giving 98% of my life date to a trillion dollar company called anthropic. Or worse Muse or OpenAI Privacy > quality difference

@xenothales @0xSero I just think he has enough gpu from all the pictures I saw and that he can plan some weeks of and use his money for that

@sudoingX Soldered, unless you have superhuman skills and still hands forget it

@WarnerTeddy Very cool !

@_lyraaaa_ 😂

@kapilansh_twt Nothing !

@p1nosaur Would be insane next to my Bambu lab p1s printer - if you have a prototype and need a guy for testing - send it to me PLEASE !

@Aditya_181105 apple, apple and another time apple there is nothing that beats it --> get a used m1 macbook pro for 750.- with 32gb ram 1tb ssd and a used base iphone 17 for 750.- --> peak life quality for 1500.-

@SidharthVijay_ everything before xai - grok is dogshit (i can say that, i had next to my claude max sub all the time a grok heavy sub and the model is mid at best - i would put it in mid chinese model tier, but all this while beeing way more expensive)

hermes, i personally dont like sharing 99% of my sensitive data to trillion dollar companys

Connor @ConnorTalksAI

2 days ago

Meta Muse or Grok Bot… What’s better for an average person? Be honest…

81 3 87 26K 17

@0xSero in my opinion from what i hear and watch and see in your yt and your x posts you are not larping you actually have the knowledge behind it ...

@Beamsters1 wait 6 months and qwen 5 flash next will bring it to us - single spark owners ...

@Youssofal_ they finally have some pressure to release fable 5.5

@0xSero in 6 months i will run that intelligence on my single spark ... patience is the key

@elshayib_ stop the engagement bait, entire x is full of it - but yea you did "not" see any of that - for those with that cooked x for you here : artificialanalysis.ai

just keep in mind - they released a model better than fable and better than every model of openai ! BUT compared to anthropic and openai they have if we checkout for example kimi k3 vs gpt 6 astra 47x times less compute used in training. FUCKING 47x times less compute - no shit they can't hold up to the user pressure

@PrashantByte09 apple definately, maybe not the best on paper but macbook, iphone, ipad and especially airpods ecosystem is peak and also apple m chip > snapdragon, intel, amd, nvidia

@Gaus450 Brave / Zai / deepseek