HPCC64

STEM || Innovations || 🐈‍⬛

[jack @jack]

we're launching BUZZ!

a new groupchat platform for teams of people and agents of all sizes, built to reduce our dependency on slack and github. model-agnostic, decentralized, self-sovereign, and open source. 🐝

https://t.co/8IaMVeTQNo

[Google @Google]

Today we’re expanding the Gemini family with three new models built to be faster, more token efficient, and reliable at scale.

Meet the new Gemini models ↓

[Tongyi Lab @Ali_TongyiLab]

Qwen-Audio-3.0-TTS is here. 🎙️

Our latest text-to-speech model, in two flavors:
• Flash: real-time interaction
• Plus: high-quality generation

What's new:
• Multilingual coverage across 16 languages
• Style control in natural language
• Fine-grained tags for non-verbal details
• More robust voice cloning from imperfect audio

[Alibaba Cloud @alibaba_cloud]

Alibaba Cloud launches its new cloud region in France!

As our third infrastructure hub in Europe, alongside Germany and the UK, the new region helps businesses scale innovation with locally hosted cloud services—and paves the way for upcoming Agentic AI services.

[Qwen @Alibaba_Qwen]

Qwen3.8 is launching and going open-weight soon!🌐

With a massive 2.4T parameters, this model is continuously evolving. We believe it’s one of the most powerful model available today, compatible to leading frontier AI models , second only to Fable 5.

You don't have to wait to test it. Just now, the Qwen3.8-Max-Preview made its debut on Alibaba’s Token Plan, Qoder, and QoderWork. Be among the very first to try it out.

Can't wait to hear what you build. Stay tuned! 🚀 

Token Plan
international：https://t.co/YRvcGdB9Bv
China：https://t.co/PKMUNwUuRp

[Jun Song @jun_song]

Guide on how to run Kimi-K3 locally :

- you need to buy 10 x GB300 (252GB VRAM each), thats about $1M
- Electricity bill estimated $10k/month

[Kimi.ai @Kimi_Moonshot]

Introducing Kimi K3: Open Frontier Intelligence

🔹 2.8 Trillion Parameters, 1 Million Context, Native Multimodal
🔹 Kimi Delta Attention enables up to 6.3x faster decoding in million-token contexts
🔹 Attention Residuals deliver ~25% higher training efficiency at <2% additional cost
🔹 Built for long-horizon agentic coding and self-evolving workflows

Kimi K3 is now live on on https://t.co/zrk6zZxZUo, Kimi Work, Kimi Code, and the Kimi API.
Open Weights by July 27, 2026.

🔗 API: https://t.co/XCrgjXAqMw
🔗 Tech blog: https://t.co/YTfiMSNM1f

[Cloudflare Changelog @CFchangelog]

R2 Data Catalog now supports read-only API tokens. You can now scope access for query engines to follow the principle of least privilege.

https://t.co/Sqn13OwSIu

[HPCC64 @hpcc64]

@RaminNasibov Tetris

[HPCC64 @hpcc64]

@leafs_s Were we ever smart at all?…

[NVIDIA AI @NVIDIAAI]

We gave a coding agent a goal and a time budget: build a training environment and teach a vision model to count colored stars.

Using autoresearch with NeMo RL, NeMo Gym, and reusable skills, the agent set up, trained and evaluated the model while the researcher steered the work.

Qwen3-VL-2B went from 25% to 96.9% accuracy, and the agent even proposed the next experiment on its own.

[HPCC64 @hpcc64]

Today’s release of OpenAI’s super app has made me consider switching AI providers for the second time — this time much more seriously — but definitely not to Claude. I’m starting to think about migrating to an open-source stack, using the Claude and OpenAI APIs only occasionally for specific tasks, and canceling my subscription entirely.

[Hoyt Emerson @HoytEmerson]

DuckDB can query a CSV sitting on a random URL like it's a local table:

SELECT * FROM read_csv('https://t.co/j714z0ghou');

No download step.

[alphaXiv @askalphaxiv]

Introducing the Global Researcher Map 🌎

We mapped every AI researcher into a visual landscape you can explore

Search your favorite authors, topics, or institutions, and see who’s behind the work

[alphaXiv @askalphaxiv]

"Gemma 4 Technical Report"

Gemma 4 is Google’s new open multimodal model family, with it being able to reason, read images, understand audio, handle long context, and run efficiently across sizes from 2.3B to 31B.

With E2B (effective 2.3B) and E4B (effective 4B), they roughly matching or beating Gemma 3 27B with about 10x fewer parameters. The audio encoder is also 78% smaller and long-context KV cache is reduced by up to 37.5%.

E4B actually beats Gemma 3 27B on 128k long-context RULER accuracy, 86.6 vs 66.0.

The 12B model didn't use separate vision and audio encoders, it actually feeds image patches and audio chunks directly into the LLM.

And for the 31B model, it is the top dense open model in Arena Text, while the 26B MoE activates only 3.8B parameters and still reaches 1438 Elo.

[Cloudflare Changelog @CFchangelog]

R2 Data Catalog now warns you before you delete objects manually. This prevents accidental corruption of your Iceberg tables.

https://t.co/d2iY9HP0eW

[Cloudflare @Cloudflare]

We are launching Workers Cache, a regionally tiered cache that sits directly in front of your Worker entrypoints. Infinitely composable, configured via standard HTTP headers. https://t.co/eBbxHIgUBA

[HPCC64 @hpcc64]

@thsottiaux Sync the customer's chats and data between ChatGPT and Codex. I need one consistent account where I can use ChatGPT for brainstorming and Codex for automating my work, instead of copying data between them.

[Math Cafe @Riazi_Cafe_en]

MIT's Advanced Data Structures
by Prof. Erik Demaine

Lecture notes: https://t.co/EBeFGVmouo

[Cloudflare @Cloudflare]

We're opening the waitlist for our Monetization Gateway, which will allow you to charge for any web page, dataset, API, or MCP tool behind Cloudflare. The charges will settle in stablecoins over the x402 open protocol. https://t.co/pvICtEIixj