Sufyan (@sfxnz)

Pinned post

GLM 5.3-Flash on 2x DGX sparks: my own EXL3 4bpw quant on stock TensorFold, no patches.

vs my NVFP4 vLLM recipe:

- 8 streams (was 2)
- 1M context (was 327k)
- 34 tok/s single, 59 aggregate (was 28 / 29)
- 2.5 min boot (was 16)

Clone, ./run.sh, done:

github.com/sfxnz/GLM-5.3-…

github.com GitHub - sfxnz/GLM-5.3-Flash-EXL3-4bpw-TensorFold-2x-DGX-Spark: GLM-5.3-Flash EXL3 4bpw on 2x DGX... GLM-5.3-Flash EXL3 4bpw on 2x DGX Spark with stock TensorFold 1.2.2: vision, thinking, MTP, 8 streams, 1M context. ./run.sh and ./stop.sh. - sfxnz/GLM-5.3-Flash-EXL3-4bpw-TensorFold-2x-DGX-Spark

Sufyan @sfxnz 59m

I don't know who's side loading apps on IOS but best to keep your devices up to date these days.

Binance @binance 7h

A vulnerability (worse_query) affecting iOS 26.7.1, 27.0.1, and 27.2 allows malicious apps to read local files from other apps, putting seed phrases, private keys, and login credentials at risk across the entire ecosystem. This is an iOS-level issue, not unique to Binance.

Immediate Action Required:
- Delete any apps installed from untrusted sources (such as sideloaded or manually signed IPAs).
- Update iOS and all apps to the latest official versions.
- If you suspect a breach, transfer funds to a new wallet created on a clean device and reset all active login sessions.

Never install unauthorized packages, and keep your software up to date.
View quoted post on X

Sufyan @sfxnz 4h

My GLM-5.3-Flash recipe on 2x DGX spark now serves 8 agents at once, each with its own long context.

• follow-up turns start in 2.6s (was 7 min)
• tool calls stream live
• Anthropic-style /v1/messages
• optional API keys

Same quality, Same speed. 1M context still works.

github.com/sfxnz/GLM-5.3-…

github.com GitHub - sfxnz/GLM-5.3-Flash-EXL3-4bpw-TensorFold-2x-DGX-Spark: GLM-5.3-Flash EXL3 4bpw on 2x DGX... GLM-5.3-Flash EXL3 4bpw on 2x DGX Spark with stock TensorFold 1.2.2: vision, thinking, MTP, 8 streams, 1M context. ./run.sh and ./stop.sh. - sfxnz/GLM-5.3-Flash-EXL3-4bpw-TensorFold-2x-DGX-Spark

Sufyan @sfxnz 4h

Didn't realise that SpaceXAI was this behind

Elon Musk Verified account @elonmusk 7h

@Dimdv99 @tetsuoai Not as good. Opus 5.5 is a masterpiece.

But Grok 5 should exceed Opus 5.5 and maybe that happens in 2 or 3 months, with continued acceleration thereafter.
View quoted post on X

Sufyan @sfxnz 7h

Per Elon Grok 4.8 will be nothing like Grok 4.7

Elon Musk Verified account @elonmusk 7h

We have dumped JAX/SGLang and rewritten the entire RL and inference stack from scratch in C/C++, which took some extra time.

Training (aka pre-training) is done. Lot of cleanup was needed on the SFT data to get that mostly right. Now, 4.8 is in RL and the evals are starting to climb. We will do a soft rollout to Bot users later this week and see how it goes.

Grok 4.8 will bear no resemblance to 4.7 (thankfully). It is not yet incredible, but, this time, the foundations are built right, so I expect it to evolve to be something special.
View quoted post on X

Sufyan

Latest posts 5