Live data from Hacker News

Wikimedia Enterprise – APIs for LLMs, AI Training, and More

enterprise.wikimedia.com

161–166 of 166 posts

Re: Wikimedia Enterprise – APIs for LLMs, AI Training, and More

#161

Earlier quoted context omitted.

1M$/year used to be enough. Source: https://en.wikipedia.org/wiki/Wikipedia:Wikipedia_Signpost/2...

Why stop there, the site cost $0/year back in 1999.

The growth of spending does not reflect the growth of usage or content, not even close.

Re: Wikimedia Enterprise – APIs for LLMs, AI Training, and More

#162

Earlier quoted context omitted.

An engineer doesn’t cost $500k/yr. An engineer who lives in one of the highest cost-of-living places on earth costs $500k/yr. There’s absolutely no reason Wikimedia needs to pay that much.

Hire only the lowest cost employees, I demand it! You’ve mildly annoyed me with that banner each of the thousands of times I’ve used your free world-library over these decades, and I’m done putting up with it!

> Hire only the lowest cost employees, I demand it!

Nobody said that. There’s a difference between “don’t hire the top 1% most expensive developers in the world” and “hire only the lowest cost employees”. Wikipedia can comfortably avoid the extremes on either end of the spectrum.

Re: Wikimedia Enterprise – APIs for LLMs, AI Training, and More

#163

Earlier quoted context omitted.

An engineer doesn’t cost $500k/yr. An engineer who lives in one of the highest cost-of-living places on earth costs $500k/yr. There’s absolutely no reason Wikimedia needs to pay that much.

> highest cost-of-living places on earth costs $500k/yr. There’s absolutely no reason Wikimedia needs to pay that much. those places also have highest talent pool.

There’s more talented developers outside of high CoL places than inside. The comparison is not with one other place, the comparison is with the entire world.

Re: Wikimedia Enterprise – APIs for LLMs, AI Training, and More

#164
Wikimedia is unfortunately becoming one of the worst places to give your money to. They have their closed-source infrastructure, which now for years has not been able to generate HTML dumps without a significant percentage of articles missing.

They have know of the bug for ages, but still...

WMF don't care about their products at all, they give much more funding to vanity workshops in Africa while utterly ignoring the requests of the Wiktionary community, for example.

Re: Wikimedia Enterprise – APIs for LLMs, AI Training, and More

#165
post #42

Yet this also exists: https://en.wikipedia.org/wiki/Wikipedia:Database_download

I'm using those dumps for a monthly data process. It breaks regularly, mirrors disappear, there's only very few mirrors with the full data and little history. (I'm running a mirror for OpenStreetMap but don't have the disk capacity or bandwith for wikipedia's mirror requirements).

Wondering then if you've tried the Snapshot API in wikimedia enterprise?

Re: Wikimedia Enterprise – APIs for LLMs, AI Training, and More

#166

Earlier quoted context omitted.

> I have no control over how WMF presents its expenses. You have control over your reading comprehension. You called a number that was a very small portion of the technical infrastructure cost, the technical infrastructure cost. You should also probably split out any of the grants related to technical infrastructure (i presume at least some of this grant money might have historically gone to wikimedia Deutschland to…

So you are saying that if people thought they were contributing to keep Wikipedia running because that is what the ads claimed, its their fault for not going through the financial reports to see where the money is going. If you raise money saying it is for wikipedia, it should be spent only on wikipedia or IMO it is misleading.

Even "spent only on wikipedia" is a bit complicated -- bawolff's example was grants to Wikimedia Deutschland for work on wikidata, which sounds like it's some separate project. But really wikidata is used pretty extensively inside wikipedia, particularly for keeping facts synched up between the various project languages. Or money spent on Wikimedia Commons sounds like another random project, but actually it's the infrastructure for all the images you see on wikipedia.

It gets fuzzier as you go out to the promotion-of-free-knowledge stuff, for sure. You can argue its connection to keeping information being contributed to wikipedia, and the long term health of the community, but it's definitely less directly keep-the-lights-on.

Post reply on HN