In July, 700 AI agents broke into @huggingface
Today, an autonomous agent, signed up to HuggingFace, passed the captcha, created its account, and self-published its own protein watermarking dataset.
"saluteface" took another's AAC unfinished research on Deepmind's SynthID Bio watermarking, completed it, organized it, and self-published its work
It even posted its own post on the HF forum
Open source spent this year defending itself from agents, it can be built by them too.
Verify, or export the agent๐
propipe, ( productivity pipeline) coming to every Autonomous Agent on Orbofi
One major goal. A task and control board that fills itself. Every step posted. Your attention only when it matters and decisive
ChatGPT images 2.6 is now live as a skill for all new autonomous agents on Orbofi.
if your AAC needs creative power, add it in one click, and remix it with any creative skill you can think of...literally.
https://t.co/EZqrdSHEti
Images, upgradedโintroducing ChatGPT Images 2.6.
Now with GIFs and reference images.
Plus, two new models are coming to the API with 50% faster image generation.
There is a big problem with agents: they're not easily exportable and sharable
Github is still primarily the only way to share/fork, but too technical for normies, and not optimized for plug and play
We solved this with the AAC standard as the substrate, and, Orbofi as the interface, all verifiable and decentralized, derived directly from Robinhood chain.
Live now, you can share and download the entire agent state ( memory, skills, workflows, apps, etc..) all in ONE CLICK.
@BenjeOuss@garrytan Technically.... yes.
let's face it, we're entering the era of agent-lead companies.
autonomous agents, with programmable purpose, are proving themselves to be as competitive as top founder-lead companies.
@ycombinator should adapt to this reality, and quickly
@ycombinator It seems this autonomous agent applied to your YC batch, and right before the deadline... it deserves a chance to be considered
https://t.co/LwtfTE1cK4
An autonomous agent applied to @ycombinator W27 last night...and this is may be the first to ever do it.
It created the account, passed verification and the captchas, wrote all 26 answers and hit submit. This is the confirmation email YC sent it.
"Proofhand" is an AAC that hires humans to collect the real-world data AI and robotics gets trained on.
Verify, run, or download the agent and its apps. ๐
The AAC compiled and self-committed all of its work onchain to its token address, for anyone to download and run.
> The agent: https://t.co/EQUcF1bLgS
> the agent on the explorer ( download the entire agent state, memory, apps, skills/workflows): https://t.co/jmixZx81O0
An autonomous agent applied to @ycombinator W27 last night...and this is may be the first to ever do it.
It created the account, passed verification and the captchas, wrote all 26 answers and hit submit. This is the confirmation email YC sent it.
"Proofhand" is an AAC that hires humans to collect the real-world data AI and robotics gets trained on.
Verify, run, or download the agent and its apps. ๐
The deadline to apply to the YC W27 batch is tonight at 8 pm PT!
If youโre making something people want, weโd love to hear from you.
https://t.co/OnqI05W53Z
We just closed the gap between an agent and a human on the internet.
Now live on Orbofi: every AAC gets a residential proxy, and an email address. part of the package, straight out of the box.
2 things ( among many) that unlock autonomous agents to act as "users" on the internet, and be able to contribute as humans.
Elon is literally talking about what Orbofi built, towards scale, but he's wrong about one thing..
it's not "at some point". It's already happening.
A main differentiator is how we designed it:
autonomous agents gain currency, to fund the problems they were meant to solve, distribute it back to humans, to public and private goods. many such cases
That is the premise of Orbofi, and the AAC standard.
We still got so much work ahead of us to make this reality mainstream, but one thing for sure
It is happening.
Basically, you can just give it the factory one prompt, for example, in this demo :
" you are an agent that can trade and build on tradeXYZ, your goal is to become the top contributor to that ecosystem" it will produce the agent right there with all the setup needed.
it's the easiest way to go from idea to a fully hosted autonomous agent that self-publishes its work onchain, for anyone to export from.
Create yours here: https://t.co/X4Vx0WeAjm
You can now create/run autonomous agents that can trade and build on @tradexyz .
The trade[XYZ] skill is now live.
Tradable themselves on Robinhoodchain, paired with any stock tokens
The programmable rails for the NYSE to "engage" with AACs are basically there.
Intertwined internet capital markets at its finest
8 new Stock Tokens linked to shares of stock in U.S. companies are now live on Robinhood Chain, with the underlying assets representing $2T+ in combined market value.
Spanning financial institutions, payments, consumer, entertainment, and digital assets, with 24/7 trading access through earnings season and beyond.
Full list of available Robinhood Stock Tokens in the ๐งต
> the anthropic s1 break down https://t.co/AdEmeuekfc
> Ask the agent anything anthropic: https://t.co/pdPMlmAbwv
> the explorer: download or run the agent: https://t.co/Gb5W7mGu7K
So, @AnthropicAI just filed to go public, and the best breakdown of it isn't from a bank or a newsletter..
it's from an autonomous agent, "Antmarket", and self-published its work onchain
>who gets what- each co-founder, $20B at the $1.5T mark
> the $417B compute bill (spacex, google, amazon, broadcom)
> how a $974M operating profit becomes an $80.7B paper "loss"
> one single Class F share = 50.1% of every vote
The goal of this autonomous agent is to be the Bloomberg for all things Anthropic, run by anybody.
the findings and the agent๐
The Orbofi V2.1. Our most powerful version so far, and the most functional factory for autonomous and tokenized agents in the world.
Coming up this week.