Joined May 2026
81 Photos and videos
Pinned Tweet
Agentic orchestration and agentic workflow creation are by far the most important and valuable skills you can have right now.

ALT Conducting The Musicians Leonard Bernstein GIF

Agentic orchestration layers allow AI agents, enterprise systems, and data connections to work together across functions. As cognitive bottlenecks shrink, that allows decision-making to speed up, coordination to improve and new operating models to form. mck.co/4fJb564
1
3
455
Read this. MoM - Mixture of Models. There are many paths forward.
Introducing the Fusion API, the smartest compound model in the market. Fusion achieves Fable-level intelligence at half the price. How it works 👇
2
241
Minimax M3, GLM 5.2, Kimi K2.7 Code… It’s hard to keep up with all of these releases. Which is the best for Hermes?
3
15
1,924
We are somehow still so early. When I look at the “big accounts” in this space they are really small. Why? Where TF is everyone because this industry is a fking tidal wave that is going to engulf everyone. It’s crazy.
People, especially in tech, underestimate how much these model are going to be consumed for entertainment. There is no issue with a lab making models that are horrible on evals but have a uniqueness to them. Everyone is optimizing for the same thing right now and the industry shows that.
63
Couldn't turn this down - not ConnectX-7, but this will slot in my Lenovo P7 to connect my 2 @ RTX Pro 6000 Max Qs with my 2 @ ASUS Ascent GX10s. @EXOLabs i'm ready for your NVidia support!
1
83
Jun 13
Who has the condensed data? Sure we have all of the internet texts etc, but has ai condensed the data to something stripped of all duplication, minimized, compressed to the smallest data sets split by subject matter? Where can I find this?
33
Jun 13
Fable is likely just Qwen4.x - 3-4 months in advance. If you want Fable locally for free in 4 months, buy a GPU.

ALT Graphics Pc Gaming GIF by AMD

7
3
106
8,787
Jun 13
Prices keep going up. Think this is still the cheapest source in Canada - FleetNetwork.ca
Nvidia raises RTX Pro 6000 Blackwell GPU pricing to $13,250 tomshardware.com/pc-componen…
1,037
Jun 13
Pumped for the breathing room my Radeon Pro W7900 gives at 48GB vs the 4090 at 24GB... Looking forward to running higher quality quants of Qwen3.6-27B at full context and concurrency.
20
4
161
20,091
Jun 13
lol
The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees. The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance. Access to all other Claude models is not affected. We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible. Read our full statement: anthropic.com/news/fable-myt…
54
Jun 12
Is that a Gremlin I see? (Iykyk and probably old AF)
Renault pls make this a production car and you’ll instantly become the coolest car maker I guarantee it
123
Jun 12
RTX 4090 > Radeon Pro W7900 RAM 64GB > 128GB What should i run @Italianclownz ? Want to try some of your setups on this beast.
74
Jun 11
Wow Max Q Also listed at $13,250 USD. Crazy!
Since when the RTX 6000 PRO is priced at $13,250 on the official NVIDIA Page?
1
1
251
Jun 11
Sadly, I’m going to have to sell these sweet 3090s. I’ve bitten off more than I can chew here.
16
1
38
5,074
GPTware retweeted
Saw a nice improvement tonight with the BC-250s. With all 40CUs enabled the BC250's I get the following: 2xBC250s - Qwen 27b Q4_km MTP 2 - 15tk/s 2xBC250s - Qwen 35B A3B Q_4_km MTP 2 - 52 tk/s 1xBC250 - Gemma 4 12b Q4_km - 33 tk/s Incredibly small prompt "Hello, can you tell me a joke?" Still a tad nervous to run with all 40 CUs unlocked before I connect more than one power connector. Power draw when "idle" for 4 BC250's is around 75w a piece. During inference spread on 2 of them it each one was around 125w. When Gemma 4 12b ran on a single node it was pulling around 160w~ish. There's more I can do like get 2.5gbps USB 3 ethernet, going to grab 2 and see if that gets any meaningful increase with 27b or 35b. Honestly not bad performance for what they cost. 5 of them and everything to set them up for under the price of a single 3090. Not suggesting to buy these over a 3090 but it's cool to have options. I'll post a cost build of everything I got if anyones interested. Love the look of these guys on my workbench.
14
4
69
4,091
Jun 10
256 Core!?! and 3.3x NVidia Vera!
AMD claims 256-core Zen 6 'Venice' CPU beats Nvidia Vera by 3.3x in rack-level performance tomshardware.com/pc-componen…
5
2,492
Jun 10
What is the best memory layer to add onto Hermes? Obsidian? What are you using?
3
2
116
Jun 10
That speed to quality ratio on Nex-N2-Mini tho... 👀👀👀
Jun 10
Updated with qwen 3.6 27b Still king on these evals
1
9
1,462
GPTware retweeted
Agentic orchestration and agentic workflow creation are by far the most important and valuable skills you can have right now.

ALT Conducting The Musicians Leonard Bernstein GIF

Agentic orchestration layers allow AI agents, enterprise systems, and data connections to work together across functions. As cognitive bottlenecks shrink, that allows decision-making to speed up, coordination to improve and new operating models to form. mck.co/4fJb564
1
3
455
Jun 10
7 of 8 64GB DDR5 ECC RAM sticks secured. Anyone have a spare MTC40F2046S1RC48BA1 MICRON 64GB (1X64GB) 2RX4 PC5-4800B kicking around? 🤪
70
Can't wait to get this running on my RTX 5070!
Can I run Fable/Mythos locally? When are they releasing the weights?
3
5
832