Loading video...
Video Failed to Load
We are moving from "Co-driving" to "Agentic" development in real time.
57,865 views • 8 months ago •via X (Twitter)
30 Comments

Listen to the full conversation between @thenickmoy, Antigravity’s research lead, and @delphina_ai.

Antigravity has been down since 5pm central, why sir?

Any help? I’m on the Ultra plan btw.

Ohhw great, I'll try it out!

Bro your tool has been down for hours, do you guys know about this?

Agent service down for more than 8 hours now, even on AI Ultra tier.

@antigravity, Agentic development signifies a shift toward more autonomous systems, and it's interesting to watch this transition.

Adding GPT-5.2-Codex to the model picker would be the ultimate test.

thanks for such a fun chat, @thenickmoy !

Agent terminated due to error You can prompt the model to try again or start a new conversation if the error persists.See our troubleshooting guide for more help. so,what are you build shit?

please readjust the token limits for Pro accounts 🙏

is it down? its not working atm for me, tried to reinstall

What happens with the service today? its very bad!!

awesome

Heyy guys, only problem is my Antigravity IDE keeps indefinitely loading the agents, it keeps stuck into "One moment, the agent is currently loading..." forever

When Google Antimatter?

great approach ! One thing to get right is making agents failing early and asking for inputs instead of ending in an infinite loop and hallucinating

@antigravity is a game changing ai Im almost done building a system that would have required 20 senior engineers and 2 years of working All thanks to @GoogleDeepMind and @antigravity teams looking forward seeing more updates

Agentic development is the new hiring frontier. Teams that hire for AI collaboration skills will dominate. This shift changes everything for tech recruitment.

Still the engineering mindset will remain fundamental even in the era of agentic development. You can't steer the agents in the right direction, if you don't have the domain knowledge.

I see policy for Pro plan was changed

Sloppity Slop

Agentic! He couldn't even properly create a Jupiter notebook and almost cost me an astronomical amount by creating an infinite loop of writing and revising text yesterday. "Luckily," he had made a mistake with the formatting, and the API rejected it!

Your Feature Skill is super effective. Normally, agents just manually edit files that were auto generated by sqlc. I always have to remind them every time they try to do that. With this skill, you just add a new rule and everything becomes smooth 👍

See our troubleshooting guide for more help."It keeps showing these prompts: "Agent terminated due to errorYou can prompt the model to try again or start a new conversation if the error persists. I would like to ask everyone: Is your Antigravity still working properly?

Co-driving made sense when LLMs were unpredictable. Now that you can reason through instructions, the time to babysit was always wasted friction. Real agentic systems compound - better architecture leads to fewer interventions needed next time.

It would be insane!

sorry 🙌🏻

Coolest time to be alive!!! 😀🙋♂️🙏💙

It’s been almost 2 months, still can’t theme on agent manager. Very annoying to use light mode at night.


