4yD

market player...

[NVIDIA AI @NVIDIAAI]

You're welcome 👊

[mr-r0b0t @mr_r0b0t]

@NVIDIAAI just gifted us a 75B MoE 🤩🤩🤩

nvidia/NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-NVFP4

75.3B total / 9.3B active compressed from Nemotron-3-Super-120B using the Iterative Puzzle framework.
1M token context support!

Perfect for your single GB10 ♥️

[4yD @4yDn5]

@Mayhem4Markets What motherboard or server chassis??

[Tom Dörr @tom_doerr]

Controls Windows GUI using LLMs without computer vision

https://t.co/qcyN9sXrHd

[Tony Simons @tonysimons_]

I made a Hermes Agent slash command cheat sheet a while back.

@NousResearch has shipped a ton of new stuff since then, so I tonight, I rebuilt it from scratch.

Every official slash command as of 7.4.2026.

Bookmark it.
Study it.
Become a HERMES LEGEND. 🤘

[Tom Dörr @tom_doerr]

Maintains identical browser fingerprints across Windows, macOS, and Linux

https://t.co/IujNww2F89

[Tom Dörr @tom_doerr]

Infrastructure as code for a Proxmox and Kubernetes homelab

https://t.co/LI5ZsRTSZK

[Tom Dörr @tom_doerr]

Curated collection of open-source quantum computing tools and libraries

https://t.co/2CD8YvZ8Xo

[Tom Dörr @tom_doerr]

Terminal interface for Longbridge trading powered by AI

https://t.co/6DZZwLwu5K

[Astro Polo @astropol0]

Google Gemini 3.5 Pro will launch with a massive 2 Million token context window by far the largest among major frontier models...

Anthropic's latest models (Fable 5, Sonnet 5, and Opus 4.8) support only up to 1 million tokens, making Gemini's rumored context window roughly 2× larger

this would be one of the largest context sizes available, allowing it to handle very large codebases, documents, and long conversations in one go. while delayed, this feature could be a major strength for Google when it launches...

[Hugging Models @HuggingModels]

Imagine running a massive GLM-5 model on consumer hardware. That's what nvidia's GLM-5.2-NVFP4 delivers with 4-bit FP4 quantization. It's a game changer for local AI, making high-end text generation accessible to more builders. #AI #MachineLearning

[Espen JD @Snixtp]

The RTX Pro 6000 has two matmul engines, and vLLM was only using one of them when I ran Qwen3.6 27B NVFP4

Marlin is nice for decoding, but CUTLASS is 3.5x faster processing large token prompts. For some reason, vLLM ran Marlin for both, because NVIDIA's checkpoint files mislabel what the model can actually do.

Fable 5 found out about all this and wrote a vLLM plugin that looks at the batch size and picks the right engine per call.

Results is almost 2x increase in prefill at all context lengths:

[h100envy @h100envy]

Alibaba engineer who leads Qwen explained the future of open agent models in 25 minutes - better than $2000 LLM training courses.

pre-train the base ->SFT -> RLHF -> tool use -> multi-modal -> ship a whole family (chat / VL / coder / math / QwQ).

That loop is why Qwen quietly became the most downloaded open model family on Hugging Face.

Qwen base + Qwen-VL + Qwen-Coder + QwQ reasoning - that's the stack.

Watch and save it, then read the article below.

[Tom Dörr @tom_doerr]

Cleans forensic traces on Linux, macOS, and Windows

https://t.co/POXh2CyEG8

[Tom Dörr @tom_doerr]

Generates realistic browser fingerprints for web scraping

https://t.co/mza4DZcnRx

[PaulineDavis @PaulineDavisLPD]

@HQNewsNow Oh, X, but obviously you mean Trump’s Co-conspirator in the 2024 election the racist immigrant from Apartheid South Africa ‘Elon Musk’ banned that ‘free speech my ass’ account.

[Rep. Anna Paulina Luna @RepLuna]

According to today’s expert witnesses: JACK RUBY AND CHARLES MANSON WERE MKULTRA ASSETS.

I am following up directly with the CIA to demand the full release of MKUltra records. The American people deserve and will be delivered the truth.

[moomoo @moomoo]

BUZZING NOW: SK hynix’s planned Nasdaq listing puts the AI memory trade back in focus, with $SKHY expected to begin trading July 10.

[Jason Luongo @JasonL_Capital]

If you're paying attention, Leopold Aschenbrenner is telling you exactly what to invest in. Power and compute are the bottleneck for AI.

10 stocks positioned for the AI buildout that he owns:

1. $TE - T1 Energy

Solar manufacturer operating a 5 GW module plant in Dallas with a second solar cell factory in Austin targeting late 2026 production. Produced 2.79 GW in 2025. Norway's Statnett assigned 50 MW of grid power to its Mo i Rana facility, positioning it as a potential AI data center hub powered by hydroelectric.

[moomoo @moomoo]

AI memory is no longer just a chip-pricing story. With Samsung and SK Hynix set to expand capacity, investors are starting to look further upstream.

The next read-through may sit in the companies behind the buildout:
Equipment, inspection, advanced packaging, testing, wafers, and specialty materials.

That's the bull case.

The risk: memory stays cyclical. If AI demand slows or capacity ramps too fast, that capex boom becomes overcapacity pressure.

The real question is no longer just who wins on memory pricing. It's who benefits from the capacity race behind it.

Learn more here
https://t.co/iyqJZiGnSS
Investing involves risk. Content provided by Moomoo Technologies Inc.