Notices by Adrian Cochrane (alcinnz@floss.social), page 8

Today I'm looking over GNU MultiPrecision's textual serialization library.

This has several public API wrappers around the core logic for different function signatures. And methodtables of underlying utilities, specified by those public APIs.

There's a replacement printf implementation repeatedly looking for "%" & examining subsequent character(s) calling the core logic where appropriate.

1/1.5!

A Webzine on ELPA packages - Philip Kaludercic: https://amodernist.com/texts/elpa-zine.html

Planning Go 1.21 Cryptography Work - Filippo Valsorda: https://words.filippo.io/dispatches/go-1-21-plan/

aria-label, aria-labelledby, and aria-describedby: What's the Difference? - Ben Myers: https://benmyers.dev/blog/aria-labels-and-descriptions/

ARIA labels and descriptions - Go Make Things: https://gomakethings.com/aria-labels-and-descriptions/

textra; A command-line application to convert images, PDFs, and audio files to text using Apple's APIs - Dylan Freedman: https://github.com/freedmand/textra (Github)
Simon Willison's summary: https://simonwillison.net/2023/Mar/23/textra/

Built on proprietary APIs, but I believe I've talked with someone here before who'd benefit from this!

PSA: LLM's are not trained on a knowledge base, they are trained on a text corpus - a collection of strings. What you get out is another collection of strings, generated using a a prompt, another text string. The output only contains "knowledge" to the extent that a human edits and curates it

The US understands the threat of foreign tech companies/apps gathering data on US citizens and manipulating their behaviour and is looking to ban TikTok.

Dear @EU_Commission, here’s an incomplete list of foreign tech companies/apps gathering data on EU citizens and manipulating their behaviour:

- Meta Platforms, Inc. (Facebook, Instagram, WhatsApp, etc.)
- Alphabet, Inc. (Google, YouTube, etc.)
- Twitter, Inc.
- Microsoft (Windows, etc.)
- Snap, Inc.

Thoughts?

#SiliconValley #BigTech #us #eu

@lightweight Meanwhile New Zealand has the lowest levels of perceived corruption...

I think "perceived" is a keyword!

@technomancy Oh, I know what Mozilla should do with GPTs: Have it write these blogposts!

Nothing of value will be lost...

/s

We did a new synchronized soundtrack for the 1927 Oswald the Lucky Rabbit short Trolley Trouble today.

It'll get uploaded one day next week, but I'm really pleased with how it came out, and excited to be sharing it.

This is the first released Oswald the Rabbit cartoon, and it was made by Walt Disney and Ub Iwerks.

It's cute, it's in the public domain, and this copy looks pretty great.

@technomancy Mozilla's very talented at nothingburgers of bloviation...

And they have a lot of money to misspend.

@technomancy Yeah, that link is vague as ever about what their actual complaints and values are! Unfortunately I can't expect anything else from the Mozilla blog...

@technomancy btw, what's this "strategy" you're responding to? Link please?

I did hear something mentioned along that line in an interview with their CEO...

@carnage4life What about when I find I prefer software standards written when I was a child, when I didn't read said standards as a child?

Also I think it counts for something if some of the films I'm nostalgic for lose their shine for me over time & others don't?

@lightweight @airshipper @cjnielsen My mind also jumps to our government's strategy of bribing Hollywood to make movies here (and undermine our own labour laws for them), then fail to grow our own industry as other countries bribe them more heavily.

It does not feel like a win to brag about making the effects for, say, X-Men First Class's climax... Compared to bragging about LotR being filmed here...

@lightweight @airshipper @cjnielsen My mind goes back to the pointless busywork of their "Action Plan for the Digital Strategy of Aotearoa" they mailed me after completely missing my recommendations.

The iwi on the otherhand seem very appreciative of data sovereignty when I offer it via Catalyst...

@airshipper you'll notice how all AWS 'deals' involve them 'generously' providing 'training'. What you don't see is that the training only for AWS' proprietary APIs & processes, and it's largely non-transferrable. It's designed to create consutants who *only* know how to provision AWS services, thereby effectively locking them (and their customers) into the AWS monoculture. MSFT is desperately trying to copy that approach with Azure. @cjnielsen

@technomancy I'm a little perplexed that GitHub has stopped working in WebKitGTK... And I'm on the latest release of my OS now with the default browser!

🤷♀️🤷

Options:

Drop focus appearance SC from #WCAG 2.2 = move to Recommendation in April 2023

Move focus appearance to AAA = re-enter CR and expected move to Rec. in June 2023

#accessibility @w3c #WebStandards

https://lists.w3.org/Archives/Public/w3c-wai-gl/2023JanMar/0342.html

@nonlinear The term "data colonialism" really seems to gel with the dynamics I'm seeing online...