🇺🇸 The Thunderbirds and Blue Angels; Air Force and Navy flying together like that doesn’t happen often.
Perfect way to open this night with a great delta formation.
Trump wanted his night, and he just had it
Writers: Lucas, Oliver
@OsintUpdates@elonmusk . Please look into this. All the major social media including X has been banned. Nepal is in dire need of international attention. BTW The protest is against the corruption and unjust for a very long time
Six months ago, We felt it in our gut.
AI agents were exploding onto the scene. Everywhere you looked, people were building tiny pockets of intelligence. But we felt something deeper… this was just the beginning.
Right then, We sensed two things:
- We needed a way for agents to work together, just as we humans do. One mind alone can be brilliant, but many minds collaborating is how we’ve built civilizations.
- We felt that agents needed memory. Not just scraps of chat history, but a central memory with fine-grained control, something any AI agent could access with the user’s permission.
We kept thinking, if AGI is coming, will it be built by a handful of people in locked rooms, or by millions of builders working together?
Think about it, the App Store changed the world because every developer could publish their creation. If only Apple made apps, it never would have scaled.
We believe the same is true for AGI. It will not come from one company or one genius, it will come from all of us.
That is why we are building Arkios
An open agent store built for privacy
Every developer able to ship powerful AI agent in days
Every user able to weave agents into workflows that run their day
We are not just building software, we are building the rails for a collective AGI.
@JakeMRuth I dont think people in general like to stay in $ABNB. We dont prefer $ABNB unless there are no options. I think we should own a product that majority of people want to use or say are happy to use.
At Karela Technologies, we recently tried something different: asking LLMs to generate the intermediary logic at runtime inside the code.
This is more than just automating code snippets. As compute becomes cheaper, these intermediary outputs start to resemble actual software end-to-end systems generated from plain language.
You won’t ask for an app. You’ll describe your goal: "I need to file my taxes."
And in response, the LLM might create a full-featured site like yourname yourtax com, ready to receive your documents, process your data, and generate filings.
Eventually, even the site might vanish. No frontend, no interface. Just the outcome you wanted.
Is this what AGI looks like in practice, @karpathy? Or are we still looking too small?