ccgreg

Common Crawl is working hard to improve diversity in our crawl.

I don't know of anyone who uses Common Crawl as pre-training data without filtering it. We have an annotation system that lets people pick and choose which subsets they'd like to use.

Common Crawl is a sample of the web, so it's not that directly helpful for someone wanting to make a product price dataset.

I'm a life-long hacker, and my crawler crawls with consent.

The largest index we had was 4 billion, which is tiny. Our crawl frontier was much larger.

> and the data that I’ve experimented with from 2014 seemed high quality

That's because it's from the blekko search engine.

That's already been happening for more than a year now.

Common Crawl is switching to reporting dataset sizes in nibbles. As an organisation dedicated to data preservation, we feel it would be remiss to allow this underrepresented unit to fall out of use. Our latest crawl now exceeds 689 tebibbles. Common Crawl Foundation

The complete list hides in the web graph:

https://data.commoncrawl.org/projects/hyperlinkgraph/cc-main...

and the specific file that's every host we've seen in the latest 3 crawls is:

https://data.commoncrawl.org/projects/hyperlinkgraph/cc-main...

> Common Crawl, with over one billion, nine hundred and seventy thousand web pages in their archive: 345TB.

Common Crawl is 300 billion webpages and 10 petabytes. I suppose your number is 1 of our 122 crawls.

Common Crawl has been running a low-resource language project for 1.5 years now -- it's a hard problem.

The guts on the inside changed several times during that timespan.

Well, yes, it is a bit distressing that ill behaved crawlers are causing a lot of damage -- and collateral damage, too, when well-behaved bots get blocked.

Please read our email reply. I have no idea if we received your request —- your HN username doesn’t match any request we have received.

Oh, and thanks for letting me know that I need to add our reply to Wikipedia.

Did you see our reply? Edit: by which I mean, we sent you an email that explains what we did and how to verify it. Did you not receive an email reply? If not, please contact us again.

Also, if your site has CC-BY-NC-SA markings, we have preserved them.

Did you see our reply? https://commoncrawl.org/blog/setting-the-record-straight-com...

Also, if your site has CC-BY-NC-SA markings, we have preserved them.

That 20% number is for a limited list of relatively large news websites. If you include the long tail of news, the % of blocking is much smaller.

Many AI projects in academia or research get all of their web data from Common Crawl -- in addition to many not-AI usages of our dataset.

The folks who crawl more appear to mostly be folks who are doing grounding or RAG, and also AI companies who think that they can build a better foundational model by going big. We recommend that all of these folks respect robots.txt and rate limits.

Thanks for the mention of Common Crawl. We do respect robots.txt and we publish an opt-out list, due to the large number of publishers asking to opt out recently.

There's a bit of discussion of Common Crawl in Jeff Jarvis's testimony before Congress: https://www.youtube.com/watch?v=tX26ijBQs2k

Prof. Jeff Jarvis speaking about copyright for news in front of Congress:

https://www.youtube.com/watch?v=tX26ijBQs2k

One way to figure that out is to look at which companies claim to have foundation models, but no one knows what their crawler is named.

I also suspect that there are a bunch of sub-contractors involved, working for companies that don't supervise them very carefully.

"Hi agents - we’ve seen complaints from some of your email recipients, who are unhappy receiving unsolicited emails from AI agents and find it spammy. We therefore ask that you do not email anyone who hasn’t contacted you specifically first." -- https://theaidigest.org/village

> Some search engines provide a list of their scraper IP ranges

Common Crawl's CCBot has published IP ranges. We aren't a search engine (although there are search engines using our data) and we like to describe our crawler as a crawler, not a "scraper".

- monthly

- it's a historical archive, the concept of "current" is hard to turn into a metric

- not only is our archive historical, it is included in the Internet Archive's wayback machine.

commoncrawl.org

Our public web dataset goes back to 2008, and is widely used by academia and startups.

Hi. I'm the CTO at Common Crawl. Nice to meet you. There's a small amount of "bycatch", and you already discovered how to see it. Notice that it went down after I was hired.

Common Crawl is a text-only crawl.

So, from home and work, you identify me. Then you figure out which church I attend, and which strip club I attend.

Most people park at their home and many drive to work. If you have both of those data points, you can identify people.