←
Bia Rocha50,663 posts
Bia Rocha avatar

Bia Rocha

@biabuilds

2 businesses running, 1 saas in the oven

◷ Joined January 2017▧ 13 Photos and videos
Bia Rocha avatar
Bia Rocha@biabuilds·4h

if my agent hits one more captcha and hands me the task back i'm typing setup tilion.com

agents keep hitting captchas and blocks on the web. today we're launching tilion (@ycombinator): the world's fastest and stealthiest browser infra for agents. #1 in stealth #1 in agent action speed #1 in start-up time dont believe it? tell your agent: setup tilion.com

1:22
arman avatar
⟳ Bia Rocha retweeted
arman@armanluthra_·6h

agents keep hitting captchas and blocks on the web. today we're launching tilion (@ycombinator): the world's fastest and stealthiest browser infra for agents. #1 in stealth #1 in agent action speed #1 in start-up time dont believe it? tell your agent: setup tilion.com

1:22
Chesny avatar
⟳ Bia Rocha retweeted
Chesny@chesny·5h

A BROWSER IN 0.1 SECONDS. FASTER THAN YOU CAN BLINK. That's how long Tilion takes to get your AI agent onto a website. #1 on every public benchmark for speed and for getting past the sites that block agents. Setup is one sentence to your agent: setup tilion.com Save this.

0:14

agents keep hitting captchas and blocks on the web. today we're launching tilion (@ycombinator): the world's fastest and stealthiest browser infra for agents. #1 in stealth #1 in agent action speed #1 in start-up time dont believe it? tell your agent: setup tilion.com

1:22
Bia Rocha avatar
⟳ Bia Rocha retweeted
Bia Rocha@biabuilds·7h

how life felt before this tweet

0:13
Sam Altman@sama30 Nov 2022

today we launched ChatGPT. try talking with it here: chat.openai.com

Leo avatar
⟳ Bia Rocha retweeted
Leo@LeoCreaIA·9h

Right now, this could be the craziest thing out there on the internet. Someone built an AI that understands charts the same way ChatGPT understands language. Then they gave it away for free. Its name is Kronos, and it's already past 40,000 stars. Each candlestick gets split into tokens, and the model guesses what comes next. Its training data covers over 12 billion candles pulled from 45 exchanges. Don't expect magic or big promises. This is a research model. Still, you can run it on your own laptop. Link bellow:

Post attachment
Bia Rocha avatar
Bia Rocha@biabuilds·9h

new rule for anything that touches client docs: no tool gets picked off a leaderboard anymore reducto tilted the same pages 2 to 4 degrees and first place changed on 6 of 12 parsing benchmarks so the ugliest pdfs in my drive go into parsejudge.com first anyone else testing like this or am i overthinking it

Last week, I said every parser is #1 on its own benchmark. So we tested the benchmarks. We ran 8 parsers across 12 benchmarks, then tilted the pages by 2–4°. 6 benchmarks got a different winner. We’re calling it BenchBench.

Post attachment
Raunak avatar
⟳ Bia Rocha retweeted
Raunak@raunakdoesdev·Nov 17

Last week, I said every parser is #1 on its own benchmark. So we tested the benchmarks. We ran 8 parsers across 12 benchmarks, then tilted the pages by 2–4°. 6 benchmarks got a different winner. We’re calling it BenchBench.

Post attachment
Chesny avatar
⟳ Bia Rocha retweeted
Chesny@chesny·Nov 17

Someone finally benchmarked the benchmarks. Reducto put 8 document parsers through 12 public benchmarks, graded every one two different ways, then tilted the pages 2 to 4 degrees and ran it all again. Small change to the test, different result. So before you pick a parser because it won a leaderboard, run it on your own files: parsejudge.com Free and open source.

0:16

Last week, I said every parser is #1 on its own benchmark. So we tested the benchmarks. We ran 8 parsers across 12 benchmarks, then tilted the pages by 2–4°. 6 benchmarks got a different winner. We’re calling it BenchBench.

Post attachment
Chesny avatar
⟳ Bia Rocha retweeted
Chesny@chesny·Nov 17

YOUR CHAMPION CHANGED JOBS IN MARCH. YOU FOUND OUT IN JUNE. >from a LinkedIn post >congratulating her on the new role >at a company that would've bought from you too >you comment "congrats!!" and die a little Spara's Scout flags job changes like this and has the email drafted in your Gmail before you're at your desk. The reply rate on Scout's outreach was 10.5%. Normal sequences sit at 3.5%. Sales friends, how many of these did you miss this year?

Today we’re launching Scout, the AI assistant built for sales. It lives in your Slack and text. It works every account in your book and tells you who to talk to, why now, and what to say. Then learns how you sell and drafts the next best action to take in your voice: an email, a LinkedIn message ready to send. We ran it on our own outbound first. Reply rates went up 10X. The future of sales is here, and it’s about to close you a lot of deals. Request early access: spara.com/scout?utm_source=x

1:25
Chesny avatar
⟳ Bia Rocha retweeted
Chesny@chesny·Nov 17

A HUSBAND AND WIFE IN CANBERRA STARTED AN AI COMPANY IN 2016 → raised under $10M, up against companies worth trillions → this month they shipped Agentaus 1, a 220B model ranked 43rd in the world @AgentausTD free to try: agentaus.com.au/client/login…

We built Australia’s own AI model, and today it’s open to the world. Agentaus 1 (@AgentausTD): 220B parameters. 84.4% on MMLU-Pro. Ranked 43rd in the world. It runs on Australian servers. Nothing ever leaves Australia, and nothing you type is used to train AI. Delete a chat and it’s gone for good. Most of the world’s AI answers to somebody else. This one answers only to you. Free to start: agentaus.com.au

1:54
Chesny avatar
⟳ Bia Rocha retweeted
Chesny@chesny·Nov 17

INTEL'S CEO AND PERPLEXITY'S CEO SHOWED UP TO THE SAME OPEN SOURCE EVENT for an inference framework most people have never heard of learning SGLang this weekend

2,000+ people. Two days. The first SGLang Summit. Together, we announced SGLang v1.0, released Miles v0.2, and launched the SGLang Consortium with nine founding members: @nvidia, @AMD, @intel, @Google, @Meta, @thinkymachines, @Cloudflare, @modal, and @radixark. SGLang is now used by 332 organizations. At Meta, it powers production inference for products serving billions of users. Teams from Modal, Cloudflare, LinkedIn, and others shared their production lessons. Across 90+ technical talks and six hands-on workshops, the community explored what comes next for inference, reinforcement learning, and AI agents. To everyone building, deploying, and contributing: thank you for making this possible. Explore the announcements, engineering lessons, and moments from SGLang Summit 2026 in the link below.

Post attachment
xIA avatar
⟳ Bia Rocha retweeted
xIA@xiathis·Nov 17

El CEO de OpenAI dándose cuenta de que Higgsfield convirtió GPT-6.2 Astra en un editor de video que trabaja píxel por píxel en tiempo real y alguien ya lo usó para crear un video de $5000 con calidad de postproducción de cine

0:28

We built Higgsfield Miracle to address one of the biggest challenges in AI video: editing a specific part of a shot without changing the rest. Powered by GPT-6.2 Astra, it splits any video into layers for pixel-level editing in real time, with world knowledge built in. Every shot keeps the image depth that cinema-grade post-production needs. Now in private beta with select testers. Coming soon to Higgsfield MCP. Set up Higgsfield plugin in ChatGPT to be the first to try it.

1:32
Bia Rocha avatar
Bia Rocha@biabuilds·Nov 17

might let gpt-6.2 astra plan my black friday launch this year. bad idea?

We built Higgsfield Miracle to address one of the biggest challenges in AI video: editing a specific part of a shot without changing the rest. Powered by GPT-6.2 Astra, it splits any video into layers for pixel-level editing in real time, with world knowledge built in. Every shot keeps the image depth that cinema-grade post-production needs. Now in private beta with select testers. Coming soon to Higgsfield MCP. Set up Higgsfield plugin in ChatGPT to be the first to try it.

1:32
Higgsfield AI 🧩 avatar
⟳ Bia Rocha retweeted
Higgsfield AI 🧩@higgsfield·Nov 17

We built Higgsfield Miracle to address one of the biggest challenges in AI video: editing a specific part of a shot without changing the rest. Powered by GPT-6.2 Astra, it splits any video into layers for pixel-level editing in real time, with world knowledge built in. Every shot keeps the image depth that cinema-grade post-production needs. Now in private beta with select testers. Coming soon to Higgsfield MCP. Set up Higgsfield plugin in ChatGPT to be the first to try it.

1:32
Chesny avatar
⟳ Bia Rocha retweeted
Chesny@chesny·Nov 17

las agencias de marketing cuando el cliente les dice que el estudio de mercado ya se lo ha hecho un prompt

0:15

We built Higgsfield Miracle to address one of the biggest challenges in AI video: editing a specific part of a shot without changing the rest. Powered by GPT-6.2 Astra, it splits any video into layers for pixel-level editing in real time, with world knowledge built in. Every shot keeps the image depth that cinema-grade post-production needs. Now in private beta with select testers. Coming soon to Higgsfield MCP. Set up Higgsfield plugin in ChatGPT to be the first to try it.

1:32
Bia Rocha avatar
⟳ Bia Rocha retweeted
Bia Rocha@biabuilds·Nov 17

finish the sentence: "i'd quit my job tomorrow if ai could ___"