You can create more advanced games with @AnthropicAI Claude even with minimal programming skills. If you know Git and can occasionally fix minor issues, that's convenient, but otherwise, it’s 95% prompt engineering. #Claude

💪 Now you can use GPT-5.2 Codex in Cursor! From the team: "We've found it to be exceptional at long-running tasks, making it great for multi-step projects that require sustained context and reasoning."

Introducing GLM-Image: A new milestone in open-source image generation. GLM-Image uses a hybrid auto-regressive plus diffusion architecture, combining strong global semantic understanding with high fidelity visual detail. It matches mainstream diffusion models in overall quality

I saw that Claude Code’s popularity has gone through the roof and people are starting to say that SaaS will end soon. It’s kind of similar to when they also said fiat currency was over. I am not an AI skeptic at all but these statements are standing on very unstable pillars.

Who could have thought? Microsoft please follow this trend.

As Sharpa Robotics says, no one saw this coming! Their robot, North, can assemble wind turbines with high precision dexterity. In the humanoid robot race, besides end to end operation, hand dexterity is critical. When you see robots doing kung fu moves and acrobatics, you can

Fast and furious

I have started playing with my really old @codewars account again and I’d forgotten how much fun it is. In the world of AI this kind of gamified platform is really useful for keeping algorithmic programming skills in shape. Do you think it will still be necessary in the future?

You’ve got only one day left to secure your Commodore 64 Ultimate at the launch year sale.

In case you missed it: Artificial Analysis now includes image and video generation benchmarks as well.

Do you remember Manus AI, which focused on building general-purpose AI agents? It has now been acquired by Meta! We know Meta is heavily investing in R&D and AI and it doesn’t seem like Zuckerberg is changing his mind anytime soon. After spending billions on data centers and

GLM-4.7 has dethroned Kimi K2 Thinking as the top open weights model on the Artificial Analysis Intelligence Index. It also placed 6th overall, very close to the frontier models.

The Commodore 64 Ultimate Founders Edition will be shipped in January if everything goes well. The new keycaps look super awesome, I think it’s worth the wait.

Finally @cursor_ai introduced a cumulative spending chart so we can easily track how close we are to the limit. In my experience, it’s very hard to reach the $20 cap, even if you’re using more expensive models.

This is the type of humor and privacy feature I really appreciate. :)

And so it begins!

First AI browsers, now IDEs…🤷‍♂️

I don’t always have suggestions but @t3dotchat 's UX is superior to its rivals.

GPT-5.2 Pro seems really slow according to OpenRouter.

Let's see if GPT-5.2 will move a needle here.

I couldn’t really vibe code a better Thruster game with matter JS using GPT-5.2. There are too many particles, the ship is ugly and there is no damage system, but at least it used the physics library properly. This is a hard task for any LLM, so I think the result is good enough.