Update thats a little past due!
2024 Recap!
Everything started picking up for me in the 2nd half of 2024! Then I lost a brand direct in Q1 of 2025 :(
That brand was a huge percentage of our revenue, but hey, life happens.
@TheAhmadOsman Ok question for you, why would I go for a setup of a bunch of 6000PROs or Nvidia GPUs (Consumer or Professional)
If I can get a M3 Ultra Mac Studio with 512GB VRAM for $10kish???
Yes I know it may be a little slower, but the energy efficiency and price makes up for it no?
@LottoLabs Very neat! I see you are running qwopus, do you like it more than Opus 4.6?? Debating whether or not I want to jump from ClaudeMax to runninf Qwopus on my 3090 locally.
For the tasks I do, I can't afford to loose any sort of accuracy (Ai Brain) or have the Ai hallucinate.
@KyleHessling1@stevibe Fantastic, will look into switching over from Hermes. A lot of the features that Hermes has me and Claude Code built (or tried to build 🤣) into OpenClaw but the constant trouble shooting is starting to get on my nerves.
@KyleHessling1@Hesamation@huggingface@danielhanchen Sweet! Gonna give v2 a go on my 3090 later today after work!
How does v2 stack up against actual frontier Models?? Opus / Sonnet 4.6 and GPT 5.4??
Does it actually work better / as good for agenic workflows like the publisher claims?
@KyleHessling1@Hesamation@huggingface@danielhanchen Hey Kyle,
I'm seeing a lot of hype around this variant of QWEN 3.5 with the Opus reasoning but people on Reddit were saying that the Opus-Distill actually performs WORSE than the normal 3.5 27B model. Can you let me know your thoughts since you've tested both V1 and V2 ?
@DavidOndrej1 You're talking out of your ass. Have you actually ran this model and compared it to the SOA frontier models? Have you compared it to Vanilla Qwen 3.5??
People on reddit are reporting its actually WORSE than 3.5 Vanilla.