2 businesses running, 1 saas in the oven
Florianópolis
Joined January 2017
- Tweets50.3K
- Following917
- Followers88K
- Likes41.9K
Bia Rocha retweeted
GROK BOT BUT FOR PEOPLE WHO ACTUALLY SHIP CODE
Easiest way I can explain Reason Machines
And it makes iPhone apps now??
Kinda want to see what it does when the first build crashes
Introducing the best Grok Bot Alternative
Meet Reason Machines
With Reasons as your teammate powered by BYOK/Sub you can code much more!
Free trial, cloud/local, and shared all in one app.
Try it for free today at reasonmachines.com
Bia Rocha retweeted
los que se pasaron 3 años aprendiendo a rotoscopiar en after effects viendo esto
Introducing Higgsfield Miracle.
Our latest proprietary post-production technology vectorizes any video into layers for localized, pixel-level editing in real time, with world knowledge built in. Every frame holds the image depth needed for cinema-grade post-production.
Now in private beta with select testers. Coming soon to Higgsfield MCP.
Set up Higgsfield MCP in Claude to be the first to try it.
Bia Rocha retweeted
Been playing w/ AI video tools for a while and editing was always the weak spot.
u generate something almost perfect and then one detail ruins it.
Higgsfield Miracle fixing that in real time, layer by layer, is exactly what was missing.
and having it land in Claude via MCP? can't wait.
Introducing Higgsfield Miracle.
Our latest proprietary post-production technology vectorizes any video into layers for localized, pixel-level editing in real time, with world knowledge built in. Every frame holds the image depth needed for cinema-grade post-production.
Now in private beta with select testers. Coming soon to Higgsfield MCP.
Set up Higgsfield MCP in Claude to be the first to try it.
Bia Rocha retweeted
every product video i've paid for had one thing i wanted to fix after delivery
and the answer was always "we'd have to reshoot"
if i can open the video and change just that layer myself, i'm not sure i book another shoot
am i wrong?
Introducing Higgsfield Miracle.
Our latest proprietary post-production technology vectorizes any video into layers for localized, pixel-level editing in real time, with world knowledge built in. Every frame holds the image depth needed for cinema-grade post-production.
Now in private beta with select testers. Coming soon to Higgsfield MCP.
Set up Higgsfield MCP in Claude to be the first to try it.
every product video i've paid for had one thing i wanted to fix after delivery
and the answer was always "we'd have to reshoot"
if i can open the video and change just that layer myself, i'm not sure i book another shoot
am i wrong?
Introducing Higgsfield Miracle.
Our latest proprietary post-production technology vectorizes any video into layers for localized, pixel-level editing in real time, with world knowledge built in. Every frame holds the image depth needed for cinema-grade post-production.
Now in private beta with select testers. Coming soon to Higgsfield MCP.
Set up Higgsfield MCP in Claude to be the first to try it.
Bia Rocha retweeted
Introducing Higgsfield Miracle.
Our latest proprietary post-production technology vectorizes any video into layers for localized, pixel-level editing in real time, with world knowledge built in. Every frame holds the image depth needed for cinema-grade post-production.
Now in private beta with select testers. Coming soon to Higgsfield MCP.
Set up Higgsfield MCP in Claude to be the first to try it.
the doc search tool every founder keeps postponing just got cheap
$1 per 1,000 pages with r-1 mini
should i finally build it?
Bia Rocha retweeted
Say a 2-person legal startup has 18,430 pages of scanned contracts sitting in a Drive folder.
Parsing all of it used to be a budget meeting.
With r-1 mini it's $18.43.
No sampling. No "we'll do the rest later." No picking which clients matter.
Reducto's whole pitch is that the tables come out with the numbers in the right rows, so the agent on top can actually answer.
$1 per 1,000 pages from @reductoai.
Save this.
Bia Rocha retweeted
Models are becoming smarter, faster, and cheaper, but the OCR industry hasn’t kept up.
We’re launching r-1 mini: the first model to offer truly premium parsing without the premium pricing. At $1/1000 pages, r-1 mini beats options that cost up to 25x as much.
Bia Rocha retweeted
HOW DO YOU GET A 5X CHEAPER HIGGSFIELD?
Voyager's answer has two parts.
Half the price per generation, on models like Seedance 2.5.
And 2-3x fewer API calls to get to a final output.
Multiply them and you land around 5x.
Curious if that second number holds up on the messy prompts, the ones that usually eat five tries.
🤝 Paid partnership
Bia Rocha retweeted
NOBODY HAD HEARD OF THIS AI COMPANY AND CHIP DESIGNERS ALREADY USE IT FOR DAILY WORK
It's called Polished Snow and it just came out of stealth.
Their first model is Case. According to David, the founder, chip design and robotics companies already use it for everyday work.
They won't say which ones, which only makes me more curious.
One line on their site stuck with me: more ideas are possible than people have time for.
We've all got one parked somewhere from years ago.
They want machines that take on the constant work so people have room left for the hard stuff.
Save plsn.ai, for now you get in by requesting access. I'll be keeping an eye on @plshdsnow.
Humanity has always found ways to do more than it once could.
Our mission is to take human progress further by solving the existential constraints that remain.
plsn.ai
Bia Rocha retweeted
A FORMER TWITTER EXEC JUST SHIPPED AI MEMORY THAT DOESN'T CARE WHICH MODEL YOU USE
Write something in ChatGPT → pull it back up in Claude
Same brain → OpenClaw, Hermes, Cursor, Codex
464ms median retrieval
Up to 22.6x fewer tokens per question vs full context
The part I keep thinking about is the 4 scoring policies. It decides what's worth keeping and when it comes back into context, so the memory doesn't just keep growing as a pile of embeddings.
I re-explain the same project every time I switch models. Curious how this holds up after a few weeks of real use.
@livingbrain
🤝 Paid partnership
Bia Rocha retweeted
In 2005 a Columbia professor started writing an open reference for modern algebraic geometry
21 years later it's 7,500+ pages. the stuff underneath Fermat's Last Theorem and Langlands
@axiommathai just turned the whole thing into Lean, about 16.5M lines a computer has to check
7,500 PAGES AND NOW IT HAS TO COMPILE
no more "the details are left to the reader" lol
the part i keep rereading is how they split it. agents organized like an eng team, some owning architecture, some the definitions, some grinding proofs
curious what people build on top of this first
🤝 Paid partnership
Bia Rocha retweeted
>one image
>10 seconds of audio
>you're on a live video call with it
AI avatars always felt stiff to me. This one actually reacts and moves like someone's there
@yuchenping @atmanity
🤝 Paid partnership
After Google acquired my AR startup, I started thinking about the next interface, this time for AI. Text and voice work fine, just like DOS did before the Mac. But nothing’s more intuitive than face to face.
After two years of research, today we’re launching Atmee v2: a real-time avatar model that makes live video agents feel truly life-like.
• Any character from 1 image + 10s of audio
• Contextual expressions, gestures, even interacting with the environment
• Live video calls, voice only, or video generation
At its core, Atmee v2 is a video foundation model, it can render anything, but it’s built to excel at characters, which is what video models like Seedance 2.5 and Wan 3.0 are mostly used for. In comparison:
Atmee v2
• Up to 50x faster
• 10-100x cheaper
Build tutors, sales coaches, support agents, companions, game characters, even whole living worlds. The interface after text and voice starts today.
Bia Rocha retweeted
IN EVERY MULTIPLAYER GAME I'VE PLAYED THE MAP IS THE ONE THING NOBODY CAN TOUCH
You can spend hundreds of hours on it and it stays exactly how the devs left it.
Parallax just dropped a world model where the map changes while you're playing.
Everyone joins the same generated world, each from their own view.
Then someone types a prompt and the whole map becomes a different one, mid-game, for all of you at once.
Knowing my friends, someone turns it into something awful within five minutes, and I want to be in that lobby.
Save this, then explore Parallax at parallaxworld.live/ and follow @parallax_live for updates.
Meet Parallax, the world's first real-time, live-editable multiplayer world model.
Share one world.
Reshape it with natural language.
See changes synced across player views.
Parallax — Worlds Without Limits
Explore Parallax: parallaxworld.live/
Follow @parallax_live for updates.