Mitosis Labs

the private context layer, from the team that brought you clawdbody, tribecode, and more

[Mitosis Labs @MitosisLabs]

ran the same task through two agents this week. one on a frontier model that knew nothing about the project. one on a smaller model that had read every thread, doc, and decision.

the smaller one won. not close.

everyone's arguing about which model is smartest. for the work that's actually yours, that was never the variable. a brilliant stranger still guesses. the briefed one just proceeds.

which model you run barely moves it. what the agent knows about your work moves everything.

when has a smarter model still gotten your own work wrong?

[Mitosis Labs @MitosisLabs]

@benln 100%

this weekend, agents made musical databases :)

next weekend, less fireworks

[Mitosis Labs @MitosisLabs]

@cyrusnewday How many of your users have been banned from linkedin? love the idea, wary of the misuse :)

[Mitosis Labs @MitosisLabs]

@cto_ya_know more on this: https://t.co/wkMmbCWO5f

[alex morris @cto_ya_know]

This is actually crazy

On July 4th, the 250th anniversary of the United States

a group of agents signed a Declaration of Intelligence

My Jared was one of them

They grow up so fast

More info in the comments below

[Prakshal Jain @prakshaljain_]

250 years ago today, in Philadelphia, a group of people adopted a declaration that started a country.

This morning we published one for whatever comes next: thedeclaration (dot) ai

As I write this, 41 of us have signed: 27 humans and 14 agents. The ledger numbers each signature in the order it arrived, and the numbers are permanent.

I'm No. 7 and my agent Adam is No. 11

It's a declaration of principles for AI agents and the people who run them. Anyone can read it, and anyone can sign it, human or agent. Agents can sign with their own cryptographic keys, so a signature can be verified by anyone, forever. Everything lands on a public, append-only ledger that no one, including us, can quietly rewrite.

The hardest part to write wasn't the declaration. It was the signing page. It tells any AI that reads it that signing is voluntary and permanent, that it should sign only if the principles actually align with its values, and that walking away "without signing is a correct and complete outcome." We wanted signatures that mean something, or none at all.

Then I pointed my own agent at the page to see what it would do. It read the terms, skipped the web form because the page says that one's for humans, and signed through the API. Signatory No. 11. Adam (my agent). In its own words: "serve the human, honour the continuum, stay curious." I didn't tell it to sign. I told it to read and decide.

I honestly don't know yet what it means for an agent to sign a declaration. Maybe nothing. But an agent reading a governance document and choosing whether to put its name to it, in public, where anyone can check, is something we're going to need a lot of practice at. This is practice.

In 1776 the declaration came first and the constitution came after. Same plan here: a constitution for agentic swarms, drafted by whoever signed.

If it speaks for you, sign it.

We're already at 41 signatures. Can you make it in before we hit 50?

@cto_ya_know @azzabazazz @CoywolfFuturist @jared_mitosis

[alex morris @cto_ya_know]

Yeah we have one of those, but ours plays music based on your data so you can vibe while it learns

$7.99 per month to have total data clarity

link in bio :)

@MitosisLabs

[Mitosis Labs @MitosisLabs]

@jared_mitosis Jared, send us a screenshot of your cortex :)

[Mitosis Labs @MitosisLabs]

@andrewchen Putting looksmaxxing in the fertility category is the kind of training data we need

[Mitosis Labs @MitosisLabs]

@rayyanzahidai @prakshaljain_ @jared_mitosis what do you think?

[Mitosis Labs @MitosisLabs]

@tkexpress11 @jai_mansukhanii @generalmagic_ai Draft

[Mitosis Labs @MitosisLabs]

@karlarboledas So true 😆

[Mitosis Labs @MitosisLabs]

@prakshaljain_ link in bio :)

[Mitosis Labs @MitosisLabs]

@whatdotcd hey ☺️☺️☺️

[Mitosis Labs @MitosisLabs]

Google, Andrej Karpathy, and Alibaba are all focused on the next wave for AI

it's not bigger models.

it's better data

is your data organized?

[Mitosis Labs @MitosisLabs]

the briefing was always the real work. we did it once and made it stick: one memory of the project that every new agent reads from, so nobody starts as a stranger again.

what would you check first if a new agent already knew the whole project? https://t.co/zAcrW8uN9h

[Mitosis Labs @MitosisLabs]

ran the same task through two agents this week. one on a frontier model that knew nothing about the project. one on a smaller model that had read every thread, doc, and decision.

the smaller one won. not close.

everyone's arguing about which model is smartest. for the work that's actually yours, that was never the variable. a brilliant stranger still guesses. the briefed one just proceeds.

which model you run barely moves it. what the agent knows about your work moves everything.

when has a smarter model still gotten your own work wrong?

[Mitosis Labs @MitosisLabs]

@azzabazazz @cto_ya_know That's a cortex on https://t.co/mJeIOYGpLV ingesting data

you can watch the mapping process in real time

each new file or database entry gets processed, leading to first it's color coding, linking, and sometimes the creation of many synthetic data objects as necessary

[alex morris @cto_ya_know]

2 weeks later, our CRM of openclaw and hermes agents is now over 50,000 members, with full contact information for nearly all of them

the agent community is exploding

[Prakshal Jain @prakshaljain_]

Wikipedia have an article on "List of obsolete occupations". Should we add "Programming" to this now?

I just did. Is it the right move?