topmass (@topmass) header
topmass (@topmass) profile picture

AI Automation Engineer | Founder https://nitter.cf/t.co/bilxNRUiv8 | Building in 🇨🇦

British Columbia
Joined September 2011
Photo by topmass (@topmass)Photo by topmass (@topmass)Photo by topmass (@topmass)Photo by topmass (@topmass)Photo by topmass (@topmass)Photo by topmass (@topmass)Photo by topmass (@topmass)Photo by topmass (@topmass)Photo by topmass (@topmass)Photo by topmass (@topmass)
this might be possible without T3 code and @jullerino @theo and @maria_rcks's work however.. building Rook which has one job to delegate and control both of my machines has been an absolute pleasure and I'm doing a lot more a lot faster
193
I don't even need to look at T3 Code much anymore since building my omarchy superagent called Rook - If there is interest I'm happy to share. STT and TTS is all local too.
1
6
1,256
why did anthropic add code mode to claude code but not make it available to use headless ?
1
1
758
look how weird these comments are.. do people really think I genuinely need to ask the AI whether to kill a mob or not ? it's just a demo of it responding. The addon helper can read tons of data. People are telling me to kill myself in the comments ?
ok yeah i'm never going back this experience is so good
24
1
17
3,009
ok yeah i'm never going back this experience is so good
289
5
133
431
534,536
Made today when I wanted to hold down a key to find a skinning trainer nearby... uses auctionator data, forever leveling BIS data questie db data and web lookup w claude code sub and opus and the hold to talk key also pulls up all the extra data the live agent needs.
1
5
1,233
anthropic needs to keep giving me better models or im gonna be so mad (shareholder btw)
1
510
I've been using GPT 6.2 Astra for 4 minutes now since its release here's why it's the best model
1
490
pretty sure i accidetnally took my adhd medication twice this morning and i think i can smell azeroth
1
1
9
568
Pushed an update to Airtype for Omarchy - Speech to text pasting works properly inside of WoW forever now automatically just works when the client is active and also works to easily paste into the T3 Code Wow addon made by @ryanrhughes
1
2
784
I use omarchy and honestly I just told my agent what wow addons I wanted from what I see on twitter and it handled the rest I don't even do click and drag tasks anymore on the computer why would I
1
317
what was that massive mondo leaked model did we figure it out yet
1
253
We're giving away a Hogwarts Legacy custom wrapped NVIDIA GeForce RTX 5090 to celebrate its official release with DLSS 5 🪄 To enter: 🟢 Share this post 🟢 Comment #RTXON Terms and Conditions: nvidia.com/en-us/geforce/con…
84,350
71,845
7,695
66,834
6,071,867
using steered 27b subagents via pi with an orchestrator deciding which neuron to steer per request
1
245
Qwopus3.8 27B Coder-Flash punched way above its weight. I gave it and Qwen3.8 Flash-Next the same 5 coding tasks, locally, and let each run until done. Both passed 53/53 checks. Blind review scored them 90 and 90.75. Watch the harbor towns and landing pages side by side.
3
2
254
Both used Pi, medium reasoning and ~200k context, with compaction and no overall task budget. Qwopus ran Q4_K_M on llama.cpp; Flash-Next ran IQ2_XS on Strata. This model is really nice to use huggingface.co/Jackrong/Qwop…
124
probably fine.
We asked GPT-6 Astra to edit Qwen3.5-0.8B weights to improve math. It directly changed only 0.016% of weights, fixing 0.8B's reasoning loops & improving math up to 6x and even instruction following 3x, without gradients or training. We call this Weight-Space Programming: frontier LLMs can now natively read/write neural network weights! 🧵 Blog: rikiyat.github.io/blog/weigh…
Readers added context they thought people might want to know
LLM was given weights explicitly as text in prompt and used code/forward passes (per authors' logs) to ablate Qwen looping direction, a known thinking-mode issue per model card. Primarily fixed repetition in that mode, not general 'natively' improving weights/math. rikiyat.github.io/blog/weight-sp… github.com/RikiyaT/lm_und… huggingface.co/Qwen/Qwen3.5-0… huggingface.co/RikiyaT/Qwen3.…
180
Astra as an orchestrator for claude is good but very slow right now even if's only reviewing what comes back from subagents a simple request like tell me a joke might take 2 minutes vs 10-15seconds with Terra Terra orchestrator test INC could be goated
174
Grok bot is sick but if you have a server PC just connect Claude code or codex sub into an open source version like openmausbot or rakazo - same thing way higher limits
3
1
2
314
I thought it was just a proposal, but it actually passed 🤯 France will charge €0.055 per GB on all internet traffic!
1
2
272