There are projects like LLM-jp: A Cross-organizational Project for the Research and Development of Fully Open Japanese LLMs looking to solve this problem by creating LLMs from open data. These are helped by a meaningful Open Source Definition, and hurt by a weak one.

A quarter century ago nobody wanted to make their source code public either; remember Ballmer: Linux is a cancer?

If you want to help ensure a future where Open Source AI models are plentiful rather than drowned out by commercial black boxes, then sign the Open Source Declaration.

Open Source protects the four essential freedoms: to use, study, modify, and share software however you want, whoever you are, and without having to ask for permission.

The OSAID doesnâ(TM)t fully protect any of them, which is why sticking with the status quo is safer: https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fosd.fyi%2F

I've had a beef with URL shorteners for a long while now for reasons that have been covered ad nauseam (not the least of which being that in addition to adding significant overhead - typically hundreds of milliseconds per request - they are just plain evil). IMO the best solution is to let webmasters create and advertise their own short links using the "shortlink" link relation (e.g. rel="shortlink" in the HTTP headers and/or HTML HEAD) such that they can be auto-detected by clients who then no longer need to generate their own using 3rd party services. I wrote the shortlink specification a few months ago (based on similar work done by others), released it into the public domain using CC Zero and went about soliciting feedback. The standard got a big shot in the arm last week when WordPress.com announced support for rel=shortlink on over 100 million pages. I've since requested support be introduced into the top 20 Twitter clients (representing over 80% of Twitter usage) and have had only positive feedback so far. A number of other high profile sites like PHP.net and Ars Technica have also jumped on board. Anyway if you, like me, are sick of URL shorteners then you're welcome to give me a hand making them go away...

Sam

I'm active in a various standards efforts including OASIS, W3C and OGF. The fact that I contributed to a charter for an interoperability working group a year ago does not preclude me from taking exception to shenanigans like this by way of a well-justified complaint.

Sam

Earlier today I created the hAl Microsoft Topic Ban incident on Wikipedia's Conflict of Interest Noticeboard, highlighting some of the particularly troubling points in the contributions of a user called hAl (who reveals little beyond liking beer). It seems I'm not the first to stumble on this apparent Microsoft shill, but hopefully I'll be the last (at least on Wikipedia) as with any luck he'll land himself a topic ban having been blocked 4 times already.

Sam

As I've explained in detail here and here, while the underlying concept is sound, the implementation has many problems:
  - "rev" is deprecated in HTML 5, so essentially a non-starter
  - "rev" and "rel" are easily confused - use the wrong one and you may well drop off the Internet
  - messing with the canonical URLs is dangerous
  - taking rather than giving canonical-ness is dangerous
  - the solution can only work for one URL (the canonical URL itself), when there can be an infinite number

A *much* better solution is to use rel="shortcut" to specify a short (but not necessarily shortest or even shorter) URL. Other alternatives like "short" are ambiguous as to whether it is the URL or its target which is "short", and "alternate shorter" are just plain wrong.

Sam

Yeah I expect the "iPad" (for want of a better name) to be more like an iPhone than a MacBook.

See Crystal Ball: Apple's $599 "iPad" Netbook (with pictures) for more.

Sam

Psion have essentially given an amnesty to bloggers and journalists using the term "netbook" (which may prove reason enough in itself to take the trademark off them since any licensing must include quality assurance). That includes blogs with advertising as explained here:

"where a blogger uses context sensitive advertising that is completely outside of its control (so it has no knowledge at all whether a 'Netbook' related advert will be placed in its blog site), then we're taking the view that we need to focus on working on persuading the featured retailer to adopt a term other than 'netbook'."

This is why we believe the amnesty doesn't go far enough.

Actually no, the netBook rather than netBook Pro figures are relevant if only because it was on the basis of a netBook flyer that Psion renewed the trademark in 2006 (long after that particular product had been discontinued). This was the basis of Dell & Intel's claims of fraud, which could well undermine the trademark altogether (assuming abandonment and/or genericide don't).