Full Context @fullctx

1 line is all it takes, Local models, Agents, The full context, building & breaking in public. github.com/thefullctx 127.0.0.1 Joined November 2010

Tweets

2K
Followers

330
Following

133
Likes

3K

Full Context @fullctx

20 minutes ago

@wikiwayne Stealth model on opencode. I think it’s minimax m3.1?

0 0 0 17 0

View Details

Full Context @fullctx

2 hours ago

I think the physics on the leaves and wind are okay but it doesnt come close to Sonnet 5.5. I used your exact prompt and made this with Space Bunny. This is one shot.

Wayne Lowry @wikiwayne

12 hours ago

3 1 6 1K 1

1 0 1 387 0

View Details

RTX Pro 6000 users rejoice! 😎 One card capped at 250W is pushing roughly 1,233 code tok/s and 882 chat tok/s across eight streams on the 27B with DFlash2 and NVFP4, using greedy decoding. Thank you @machinegenie for giving us access to see what TensorFold can do. 0.6.2 has even more to give! 🤯 Also we will be we adding support for 3090’s.

15 7 98 5K 33

View Details

Full Context @fullctx

4 hours ago

@Brjen Wow, this is fire 🔥🔥🔥

1 0 0 5 0

View Details

Full Context @fullctx

4 hours ago

@wikiwayne Let me try this prompt with a different model.

0 0 0 13 0

View Details

Wayne Lowry @wikiwayne

12 hours ago

I gave Sonnet 5.5 a hard one: build an autumn foliage simulator in a single HTML file. Layered paper-cutout look. Big tree, rolling hills, grass. Rough-edged paper, grain, soft shadows. 1,500+ leaves that flutter down (not straight down) and pile up. Drag a finger to make a gust that blows falling and settled leaves and shakes the branches. Wind slider. Regrow button. 60 fps on iPhone. The result: a 23 KB file that ran on the first try. The leaves did everything I asked. It looks stellar. Totally nailed it. Setup: I ran it through the Nous Research portal, with my Hermes agent as the orchestrator. Full prompt and in the first comment.

3 1 6 1K 1

View Details

Full Context @fullctx

4 hours ago

@HermesWatcher I can confirm, i have just updated to latest version. Its way faster than it was.

0 0 0 37 0

View Details

Full Context @fullctx

4 hours ago

@MichaelGannotti Let us now how its working out for you. We can’t keep up anymore 😅

0 0 0 5 0

View Details

Full Context @fullctx

4 hours ago

@tylermayberry @MichaelGannotti Hahah, same here. I always thought 16GB was enough a couple of years ago. Now I'm on 48GB RAM on my MacBook and still think I should have gone with more.

0 0 0 6 0

View Details

Full Context @fullctx

4 hours ago

Now I really want to see these three tested on the same real-world agent tasks: Jev vs Clef vs Laya. Same hardware. Same tasks. Same conditions.

0 0 0 10 0

View Details

Full Context @fullctx

4 hours ago

Cloudflare just dropped Clef! If you’ve been following the whole Jev / Laya decision-model thing, this is definitely worth checking out. Here’s a quick breakdown 🧵 huggingface.co/Cloudflare/clef

5 0 3 134 0

View Details

Full Context @fullctx

4 hours ago

On the same benchmarks, Cloudflare reports Clef substantially ahead of Laya on several quality metrics. So the tradeoff is pretty interesting: Clef → higher reported decision quality Laya → dramatically lower latency And Clef is open-weight under Apache 2.0.

0 0 0 9 0

View Details

Full Context @fullctx

4 hours ago

Cloudflare’s Decision Index reports these median latencies: Clef: 209 ms Jev: 524 ms Laya: 5.8 ms 🤯 That Laya number is absolutely wild. But speed isn’t the whole story.

0 0 0 16 0

View Details

Full Context @fullctx

4 hours ago

Clef is API-compatible with Jev/SystemOne. So if you’re already building around that interface, experimenting with Clef should be relatively straightforward.

0 0 0 10 0

View Details

Full Context @fullctx

4 hours ago

Clef is a 27B multimodal decision model built on Qwen3.8-27B. Instead of generating text, it takes a state + typed questions and returns probabilities for the allowed answers. No free-form text. No output parsing.

0 0 0 15 0

View Details

Full Context @fullctx

5 hours ago

It’s blowing up! The community is growing, the support is growing, and you’re absolutely right about mastering one thing before trying to master a couple things mediocre. “I fear not the man who has practiced 10,000 kicks once, but I fear the man who has practiced one kick 10,000 times.” Bruce Lee 🥋

1 0 1 6 0

View Details

Full Context @fullctx

5 hours ago

TensorFold flew under my radar, and it shouldn’t have. I’ve been on vLLM for a while and if you’re running local models on a DGX Spark, this one is worth knowing about. It’s from @ashxhart, who spent months on a different idea of inference: don’t treat the whole model like it has to live in memory all the time. Stream the weights, keep the resident footprint smaller, and spend that memory on context. There’s an MLX path for Apple Silicon and a CUDA path for Spark. The CUDA side is narrow on purpose. Hand-written kernels for Qwen3.8-27B, Qwen3.8 Flash Next, and GLM-5.3-Flash, built around speculative decoding and CUDA graphs.

13 0 8 492 1

View Details

Full Context @fullctx

5 hours ago

@0xVRAM @nvidia @nvidiaAI That is a first hahahah

0 0 0 5 0

View Details

Full Context @fullctx

22 hours ago

Look at the current prices of NVIDIA DGX Spark. $6500 for the 1 TB. Are we going to see $10K?

4 0 8 569 0

View Details

Full Context @fullctx

5 hours ago

@jayleaton @landontgreen 😂😂😂 i actually trying to learn more about benchmarks. But its at a slow pace, but i want to cover benchmarks too in the near future.

1 0 0 7 0

View Details

Full Context @fullctx

5 hours ago

@thybead @ashxhart Not a second pass. Long context is where the speed falls off, because every step has to read a much bigger history. Nobody has published a number yet for the check by itself.

0 0 1 3 0

View Details

Full Context @fullctx

5 hours ago

@jayleaton @landontgreen I have just recently found out about tensorfold 👀, didnt run any benches yet. But it feels solid and consistent so far. Will soon follow up with real tests.