amrrs

that's a very interesting point to test new models. I speak an Indian language called "Tamil" and have always tested new models with Tamil but also have tried little bit of Arabic (quranic verses) with previous GPT image 2 and Nanobanana pro and they have nailed it. Don't know if it was because of extensive training data.

As someone who's trying to get better assessments, I'm struggling to come up with objective coding tasks that evaluates all aspects of real life like planning, design choices, problem solving and context usage. From your experience with humans, Do you have any recommendations on what could be effective in measuring it?

Siri AI - looks like finally Siri is getting its due update, hopefully they ship it soon!

Honestly the problem with these is how empirical it is, how someone can reproduce this? I love when Labs go beyond traditional benchies like MMLU and friends but these kind of statements don't help much either - unless it's a proper controlled study!

is that str.replace(g,t) ?

Why are you religiously defending Israel?

And you already bet this story is a fabrication as well.

This is exactly who media takes advantage of not the one who waits for investigation and acts rationally.

If going by your recent comments, I can say I bet you're just an Israeli propagandist. Would you be happy with that assesment?

Have you tried the new GLM 4.7?

Thanks for sharing your repo..looks super cool.. I'm planning to try out. Is it based on mlx or just hf transformers?

Not even a joke anymore

https://x.com/fidjissimo/status/2011592010881446116?s=20

Unethical conduct is negotiating with Zuck? :D jokes aside. but it must something serious enough for them to part ways with the co-founder

On fal, it takes less than a second many times.

https://fal.ai/models/fal-ai/z-image/turbo/api

Couple that with the LoRA, in about 3 seconds you can generate completely personalized images.

The speed alone is a big factor but if you put the model side by side with seedream and nanobanana and other models it's definitely in the top 5 and that's killer combo imho.

Honestly speaking Netflix has good catalog, much more comparable to Hollywood. Take the latest Frankenstein for example.

Don't look at only series. They also have recipes repurposed. But they acquire good titles and also produce some good ones.

If at all anything, Claude Code's success disproved this

It's built on Chromium - https://help.openai.com/en/articles/12628461-setting-up-the-...

I think it's more like Apple - use ChatGPT as the iPhone and build an ecosystem around it

It's heartbreaking and sad. His video a couple of days back was literally titled "You thought I was gone! Speedrun retruns"

The comments on that video was so kind and heartwarming where people wished him well.

While we don't know the exact cause, we can all agree that he was subjected to extreme bullying and no one stood up for him - most importantly FIDE!

Are you running the bot with the free tier api?

It is a new version of GPT-5 that's been primarily optimized for coding. Hence this confusing name - GPT-5-Codex.

This model is available inside all OpenAI codex products. Yet to be available on Api

The model is supposed to be better at code reviews and Comments than the other GPT-5 variant. It can also think/work upto 7 hours.

Surprising I don't think Atlassian hasn't even recovered from Loom acquisition. It was good for Loom but hardly any positive impact for Atlassian. Now Arc and Dia - it seems to be an obsession to acquire an AI BROWSER

Any example or prompt you use to make this statment?

Google AI Studio allows you to branch from a point in any conversation

Gemma 3 270M running on a midrange pixel device - https://x.com/1littlecoder/status/1956065040563331344

Noob Q: How do they know it's an interstellar comet? With the speed of movement between two frames?

Is anyone here after PewDiePie comment ?

Google Gemini Live is pretty good.

If you want to try only voice, Try unmute.sh by Kyutai which will be eventually open-sourced

System Card https://www-cdn.anthropic.com/6be99a52cb68eb70eb9572b4cafad1... (Search FDA)

It's thinking process to go about guessing a place is further fascinating. Even o4 mini high is quite good[1] and very fast.

But unlike a geogussr, it uses websearch[1] [1] https://youtu.be/P2QB-fpZlFk?si=7dwlTHsV_a0kHyMl [1]

The entire licensing is such a mess and Mark Zuckerberg still thinks Llama 4 is open source!

no commercial usage above 700M MAU

prefix "llama" in any redistribution eg: fine-tuning

mention "built with llama"

add license notice in all redistribution

It still answers Elon Musk as the biggest misinformation spreader - https://x.com/i/grok/share/5N2eKM8sRiaCQB6eOoYZUUwIv

It's funny they don't have Hugging Face as their Partner. Literally, the biggest face of Open LLMs sitting right in Europe, but somehow it's not a partner.