
A practice that only ever has goals is a job. A practice that never has any goals is a hobby. The good work lives in between.

Pro rider in the uncanny valley.

A practice that only ever has goals is a job. A practice that never has any goals is a hobby. The good work lives in between.

I feel like, AI like money, is not a great equilizer, but a great amplifier.

So all this announcement to say that we offer low ttft, just that? save those GPUs, train some good models, my bois!
Today, we are launching Sarvam Token Factory for everyone. We host our own models, and the full range of best-in-class open-weight models, in India. We match leading inference providers on performance and reliability. For enterprises, this means more models to choose from, and an AI stack they can own without lock-ins.

Just tested it, and why it's so good at frontend, it just doesn't make sense.
Hy4 flash - the CHEAPEST & FASTEST model USEFUL for MOST everyday tasks! RMB 0.2 / USD 0.03 per TASK --- 1/20 Hy4, 1/8 Ds4.1! Try it FREE on WorkBuddy & OpenCode, limited 2 weeks Let me know how you like the model, I'll reply personally!

But deepseek beats it on pricing, if somehow @Baidu_Inc can make it on par with deepseek pricing, lo and behold the havoc it wrecks. 3/3

It found lot of flaws in @deepseek_ai written spec files, and deepseek was writing genuine praises in his favour, being almost half in size and punching in ~800B category is really impressive. 2/3

Ernie 5.5 flash is sooo good by @Baidu_Inc , have been using it for a while now, and it totally destroys deepseek v4.1 flash (latest). I know what i am saying, but it's true, it somehow gives opus vibes in nature, like calm and trustworthy. 1/3



my chrome vulnerability was just awarded the highest bounty in @GoogleVRP history - $1,333,337


Why my 5-hour limit is burning like crazy, any idea @lydiahallie @addyosmani all this just at 141k context.

Just downloaded @t3dotcodes and learned that it just works with a third-party subscription, i used to think that you can just whip it an api-key and base-url and good to go.

anyone who wants to buy, i can supply it at $40.
3gb gddr7 is like 100$ per module, upgrade only makes sense in China where you can’t get the 6000 pro, OR if you already have the 5090 and want to upgrade it. More info over the next few months. I think this is why 5090 prices shot to 7k each

So what happened to that openAI keyboard kinda thing, are we still using it?

Opened gemini after a long time to try the new Argon model, and mind you i have the pro subscription. Sent 5 msgs in that chat, not even coding with it, and it says that i am out of current usage. what kind of usage is this? are you guys joking or what? cc @OfficialLoganK @GeminiApp