Chaz

AI and Wireless Engineer. Citizen of the world.

[Chaz @SportyKurd]

@OpenAIDevs Nice and quiet room, crystal clear channel, and where is the server oh maybe in the next room .. right! Great testing scenario :/

[Chaz @SportyKurd]

@peer_rich and they ask why is EU failing to compete -- they can't run another instance of the service .. in the age of agents

[Chaz @SportyKurd]

Insane .. go BABA! #alibaba

[Qwen @Alibaba_Qwen]

🚀 Qwen3.5-397B-A17B is here: The first open-weight model in the Qwen3.5 series.

🖼️Native multimodal. Trained for real-world agents.
✨Powered by hybrid linear attention + sparse MoE and large-scale RL environment scaling.
⚡8.6x–19.0x decoding throughput vs Qwen3-Max
🌍201 languages & dialects
📜Apache2.0 licensed

🔗Dive in:
GitHub: https://t.co/NzNdS9joAT
Chat: https://t.co/bg4tAU0Rhw
API：https://t.co/YiiyKTnHoU
Qwen Code: https://t.co/qqwj5nAger
Hugging Face: https://t.co/wFMdX5p5um
ModelScope: https://t.co/9NGXcId57a
blog: https://t.co/AW8UQStXaL

[Chaz @SportyKurd]

Can't believe Sparse Auto-encoders, which I was researching exactly 10 years ago during my PhD (and one of my favourite ML algorithms) is back in the scene as very useful tool. This will be big in the near future.

[Google DeepMind @GoogleDeepMind]

AI systems can be powerful but opaque "black boxes" - even to researchers who train them. ⬛

Enter Gemma Scope: a set of open tools made up of sparse autoencoders to help decode the inner workings of Gemma 2 models, and better address safety issues. → https://t.co/kIVwMgVwtY

[Chaz @SportyKurd]

@CALLAGHANStuart @passportwaiting @SimonCalder Hey, I have the same problem just now. How did you solve it? hope to hear back from soon

[Chaz @SportyKurd]

@Adobe What a complete rip-off you are. They say it's free to try, then commits you to an annual contract after a week!! we should boycott such companies. 🚫ADOBE🚫

[Chaz @SportyKurd]

It was a pleasure to lead the Science and Innovation stream of @kurdishhouselon's latest Employability Series networking event. There was an overwhelming number participants interested in AI. I encourage everyone under 30 y.o. to checkout opportunities at https://t.co/DIdn5joELW.

[Kurdish House London @kurdishhouselon]

Our 'science & innovation' group was the most popular group in our Employability Series networking event yesterday. Thank-you @SportyKurd and @FanyaIsmail ! Are you are an innovator in the UK under the age of 30? Check this out. https://t.co/yZl1DQuvbE @ZeyBinboga

[Kurdish House London @kurdishhouselon]

Sign up now for what could be a life changing online event this Saturday 15th May at 5.30pm-7pm. Meet with Kurdish professionals & entrepreneurs in the UK & elsewhere to discuss what fulfilling work could look like for you. Please join us. #TwitterKurds https://t.co/syDVywnkbA

[Chaz @SportyKurd]

https://t.co/gFx9YpzXQa Join us for a unique opportunity to connect online with Kurdish entrepreneurs and industry practitioners across areas from farming to business building to the creative arts industries. 15/05/2021

[Chaz @SportyKurd]

Hello, I am a noble knight

[Rawaz Rauf @rawaz]

Hello, I am a cairn in the middle and top of a mountain ....

Source : https://t.co/c9yukxXRiiڕەوەز&f=0&d=2

[Chaz @SportyKurd]

@tre_abubakr @shenemohammed So beautiful and full of meaning

[Chaz @SportyKurd]

@fchollet And why do you think it’s so useful?

[Chaz @SportyKurd]

#NewProfilePic

[Anonymous @YourAnonNews]

Jin, Jiyan, Azadi

[Google AI @GoogleAI]

Introducing TracIn, a simple, scalable approach to understand the influence of training data on #MachineLearning models by tracing the changes in prediction as individual training examples are visited during the training process. Learn all about it below: https://t.co/2HFtiKeQbT

[Chaz @SportyKurd]

A new architecture – Chipmaking is being redesigned. Effects will be far-reaching https://t.co/tSi5njgNBD

[Chaz @SportyKurd]

@Herormo Haha hope that’s not true, I am worried about the consequences 😅

[Chaz @SportyKurd]

@pjreddie @JeffDean @GoogleAI I guess he is going to say that Google's peer review is more stringent xD and finally caught one AI Ethicists paper. What he doesn't emphasise in his letter that it was Google's algos that were scrutinised .. now I wonder how many such google-negative paper pass their review!?

[Chaz @SportyKurd]

Fab

[Christian Szegedy @ChrSzegedy]

Interesting paper: 94% on cifar-10 with 80 labeled examples (8/class).

[Chaz @SportyKurd]

@LawenAzad Get better soon :)