Web3 Content Creator
Joined September 2011
- Tweets19.5K
- Following3K
- Followers9.7K
- Likes18.9K
Vincent retweeted
Your AI studio. Your servers. Your rules.
Bring your brand and your API keys.
Self-host Genfeed and make it yours.
Go make some good sh*t.
youtu.be/kH3jy7CK1YU
We’re open-sourcing the best AI studio in the world: @GenfeedAI
Images, videos, posts. Your brand, your content.
Now you can self-host the whole damn thing.
Fork it. Run it on your own servers. Make it yours.
github.com/genfeedai/genfeed…
Vincent retweeted
Introducing Genfeed.ai
The open-source content engine your AI agent can run.
Video, images and posts, from brief to publish.
Vincent retweeted
Months of building Genfeed. Today it’s open source.
An agentic AI Studio that survives the next model, built around context and a harness that learns what performs.
Now, getting discovered is the hard part.
So I’m giving the repo to as many builders as I can, and I’ll build the distribution playbook in public.
Introducing Genfeed.ai
The open-source content engine your AI agent can run.
Video, images and posts, from brief to publish.
Vincent retweeted
AI video models are cracked. The generation isn't.
HeyGen just released its video model.
MiniMax H3 underneath, post-trained by HeyGen.
A 30-second clip at $0.01/s is $0.30.
Nobody goes broke on one generation.
You go broke on the retries.
The clip that's off-brand.
The wrong voice.
The one that's fine but you'd never post.
Every one of those is credits in the trash, and they add up way faster than the price per second says.
Three things decide if a generation gets posted or trashed.
The context
A model that doesn't know your brand, your audience, what you posted last week and what flopped gives you something generic. So you regenerate. Then again.
Context-aware means brand, voice, past posts and platform go in before the first generation, not after the fifth retry.
The harness
The model is one piece. The harness is everything around it: which model for which job, the prompt, the references, the checks before a human sees anything, what happens when it fails.
HeyGen just showed it.
Same base model anyone can call, post-trained and wrapped for one job.
Same model, bad harness, you get slop.
Good harness, you get something you'd actually post.
The analytics
Code has tests. Content has nothing until you publish.
Then you get the real signal: views, replies, follows, clicks. That's the reward for the agent.
Most tools generate, publish and forget, so tomorrow's batch is as blind as today's.
Post-publish analytics are the RL for content.
What worked goes back into the context.
What flopped gets killed.
Every post should make the next one cheaper.
We still need to crack better context, better harness, better RL.
We're releasing HeyGen Video, built for businesses that need production-quality video without production-level costs.
Pricing starts at $0.01/s through October (50% off)
Built on @Minimax_AI H3, post-trained by HeyGen.
Learn more: developers.heygen.com/heygen…
Vincent retweeted
With Fable 5.1 and GPT Astra out, SaaS needs a new business model.
Seats and training aren’t enough. If you don’t price the output, you’re still selling hope.
AGI killed the pure subscription. Align to performance or die slow.
Vincent retweeted
🔴 Claude Fable 5 is OUT!
Last week, we one-shot a FPS.
Today, we are building an AI-native Game studio.
Going LIVE NOW!
youtube.com/watch?v=Ap5vza8q…
Vincent retweeted
ChatGPT Images 2.0 is the first image launch in a while that actually feels useful: posters, layouts, mockups, typography, real assets.
We’re breaking down the best prompts, examples, and reactions live in 20 minutes.
See you in chat
youtube.com/watch?v=b9Re90K4…
🤖 Made with AI
Vincent retweeted
Yo VCs, we are actively raising money.
(PD: It's a love story)
Context:
I'm an entrepreneur. I've built @smithiitools and scaled it to $2M in revenue, but it's my first time raising.
> We are building Lince, a Consumer App Targeting Europeans
Why?
In December 2024, when I was buying my engagement ring, I spent hours talking with the owner of the jewelry store.
Long story short:
He was like, " What is a 20-year-old buying a diamond for a proposal?", so I pitched him DeFi.
He ended up asking me to manage part of his money in DeFi to show him how it works.
> That was my aha moment.
> Most Europeans keep their savings in accounts yielding under 0.5%, well below inflation.
And DeFi has proven it can generate consistent yield, but it remains inaccessible to mainstream users. Also, there are a few EURC opportunities.
> Lince bridges this gap.
Users deposit euros and receive a seamless savings experience.
> The core of the product is our Smart Risk Engine: a hybrid system that continuously monitors Solana DeFi protocols and rebalances capital in real time based on yield opportunities and risk signals, optimizing for capital preservation and stability, not just maximum APY.
> We've already validated the Smart Risk Engine through our existing Solana DeFi yield product, which has +2,000 users and $1M in deposits.
> Europe = Regulations.
The compliance layer is built on regulated infrastructure from day one, while we validate Product Market Fit: Bridge handles fiat rails, KYC, and IBAN connectivity; Privy manages non-custodial wallets and user authentication. Then, regulation through an EMI License comes as the next natural step once traction is proven.
> Our only team milestone is not just Smithii.
We also count on a Platform Engineer with experience building infrastructure for banks and insurance companies.
Also, the jewelry owner is part of the team, not because of the story, but because he is also an investor and participates in several startups, having a deep network in the Spanish Ecosystem.
> Deck below.
If you are a VC actively deploying capital and Lince sounds interesting to you, feel free to DM, so we can book a call next week.
If you are not a VC, you can rather introduce a VC to us or just give feedback to our deck (much appreciated)
The shift in the crypto fundraising landscape the past 6 months has been insane.
Crypto VCs used to have to constantly be networking/writing/podcasting/going on spaces/promoting your thesis/getting on 10 deal flow calls a week, to get into good deals...now it's literally enough to just have capital to write checks.
Deals are being pushed rather than dug out. Inbound if people know you have money is at an all-time high.
Most firms are either 1) Out of money 2) Moved to Series A and beyond or 3) Fundraising (with no success).
Deals that used to close in 2-3 weeks now close in 2-3 months.
Firms with questionable business models or copy pasta of the latest trend are getting zero primary or follow-on funding (Good news!).
There are now realistically <20 firms writing checks in pre-seed/seed.
VCs basically have the pick of any deal they want, with more time to do DD.
IMHO 25/26 are going to be historic vintages for those who stick around.
Vincent retweeted
Paperclip lets you hire AI agents and run an entire company with zero human employees.
We tested it live. Built an AI CEO, CTO, and engineering team from scratch.
Gave them one mission: rebuild Sora as open source.
They made architecture decisions. Chose the tech stack. Started building the MVP. All on their own.
Part 2 coming up soon.
Vincent retweeted
We just open-sourced Paperclip: the orchestration layer for zero-human companies
It's everything you need to run an autonomous business: org charts, goal alignment, task ownership, budgets, agent templates
Just run `npx paperclipai onboard`
github.com/paperclipai/paper…
More 👇
Vincent retweeted
We're going live in one hour to test Paperclip.
Hire AI agents. Set goals.
Let them build your entire product.
The challenge:
rebuild @soraofficialapp from scratch in 60 minutes.
Zero human employees.
Just AI agents running everything.
OpenAI shut it down. We're bringing it back.
Vincent retweeted
Anthropic shipped Claude Code Channels and now, you can work from Discord or even Telegram.
Sound familiar? That's what OpenClaw has been doing for months.
We set it up live. Demoed both side by side.
One is a remote control. The other is an operating system.
Full breakdown with @mntll_nl
