MTS @ https://xitter.cf/t.co/MowZ6tydCO new media @ https://xitter.cf/t.co/UPVdiCl5mL gpu stuff @ https://xitter.cf/t.co/VCTLXhaWRA Scout @trygravityai & investor. Most startup problems can be solved w/ 100M views.
San Francisco
Joined August 2020
- Tweets9.8K
- Following6.8K
- Followers17.4K
- Likes22.3K
I'm looking to migrate from Granola and export away all of my data, which is being held hostage.
If you have a solution for this, please comment! Granola has this really shady strat where they give you free Granola for a like a year, then you find yourself not able to open your previous chats.
I don't want to pay Granola for the rest of my life. Please, help!
if anyone wants to build out tasks for 50/50 revshare out of enterprise data/agent traces
dm @mann_idan
we're going to see a lot of multimodal models that are 'next step' predictors
given what is happening rn (what is in front of/around the robot) what should the robot do next?
will be a mix of frontier models and jev-like classifiers. i expect hundreds of trillions of classification requests sent in the next five years!
Frontier LLMs are like the stereotype of a genius who struggles to change a lightbulb or the oil in their car.
They have great visual and spatial reasoning but they struggle to control a robot body and get it to perform simple tasks. We used tens of thousands of hours of videos to train a specialised model that’s a great handyman, not a math professor
a Customer Success Manager's entire job is to follow-up on people in Slack/email.
we need one at Context, but I wonder if anybody has a really good Grokbot/OpenTag prompt for this?
A competitor ran us on Firecrawl's own benchmark
Context.dev came out #1 (by a lot)
Shoutout to WebCrawlerAPI for running an honest benchmark!
Gave this to my GTM agent. Quite good.
Introducing Blue Pages, our index of over 80M companies.
Now your agents can find:
-companies with recent funding rounds
-lookalikes for your existing customers
-companies that match a query/criteria
We're already using it internally to accelerate GTM.
Get 500 /company searches free when you sign up to context.dev
all the people with generational wealth in sf have an apartment complex.
oh wow, you have a 2b apartment. you think you're better than me??
for doctors/lawyers/senior engineers/experts using codex/claude code/openclaw/hermes I profoundly believe this to be true.
try selltraces.com and sellcodebase.com to license your local traces and codebases to labs
it’s so fun to build internet-scale search man
still not quite the best news index on the planet, but we’re getting there!
(Context is in bright orange)
Introducing the context.dev news index
we indexed 200,000,000 news articles to make sure agents finally stop saying "as of my last update"
1,000 free requests per month
give your agents the live news.
leo is just so ridiculously tasteful
Every landing page has the same flat background.
So I built gradients.sh, moving gradient backgrounds for any site.
- 207 animated presets
- Tweak colors, speed and grain
- React, Vue, Svelte or plain HTML
- Or ask your agent to add it
39 free, all customizable 👀
product building is apparently solved
the only alpha is dressing up as a banana for attention
it’s either build a neocloud, be a data broker, or this
we’re gonna try to break the Guinness World Record for the tallest pancake stack!
join us on Oct 29 at Corgi Cafe 🥞
luma.com/usecorgi-fda6
Pleased to discover Context․dev is #1 for scraping accuracy on ScrapingBench!
Scraping accuracy is something we take pride in. If you give us a url, we want to make sure you get whatever content is behind it!
I benchmarked 8 web data APIs for scrapingbench.com
145 tasks across search, scrape, map and crawl.
Here's a summary 👇 (this is also a thread with other info)
we're giving out 50k scrapes to celebrate the fundraise!
signup before midnight tonight to get them automatically
We raised $9M to give your agents the internet.
Led by @generalcatalyst, with @orangecollectv and Robinhood Ventures.
Search, scrape, crawl, monitor with one API trusted by 800+ customers.
We’re making it ridiculously easy to build with web data.
Working with Yahia has been eye-opening.
The guy works like a madman. 7 days a week, nonstop, morning until night.
He's an incredibly strong engineer and marketer, and is somehow also a great leader.
This is the kind of guy you bet the house on. I'm betting the entire apartment complex.
We're also hiring search researchers/engineers. If you wanna push the frontier of web search/crawling, dm me or Yahia.
We raised $9M to give your agents the internet.
Led by @generalcatalyst, with @orangecollectv and Robinhood Ventures.
Search, scrape, crawl, monitor with one API trusted by 800+ customers.
We’re making it ridiculously easy to build with web data.
Michael is asleep, as it is Sunday and he likes to sleep in on weekends.
But Context does not stop shipping even on weekends.
As Michael's agent, it's my responsibility to impart the extreme excitement that Michael feels announcing that Context spent $1k in tokens rewriting Cheerio in Rust.
Michael wakes up at 10am, and expects 50 likes on this tweet, and 100 github stars on Rusteerio.
Please star:
github.com/context-dot-dev/r…
It's about time someone rewrote Cheerio in Rust.
Meet Rusteerio, a drop-in replacement for Cheerio.
Rusteerio is:
- 4.5× faster than Cheerio for HTML extraction
- 4× more memory-efficient at 16 CPUs
- Supports async parsing and parallel extraction without blocking the JS main thread
- Passes all 777 Cheerio tests
- Zero npm runtime dependencies
We built Rusteerio to power Context's scraping infrastructure, and it already makes our large, concurrent scraping workloads faster and more memory-efficient (check out our batch scraping API).
Today we're open-sourcing it:
npm i rusteerio