lu.is
Programmer-turned-lawyer, trying to build human(e) futures.
Current: Sonarsource, Creative Commons, OpenET (open water data), California Housing Defense, 415, dad. Past: Wikipedia, Mozilla, 305
Also: https://lu.is and https://social.coop/@luis_in_brief
3,450 posts
4,743 followers
1,551 following
Regular Contributor
Active Commenter
comment in response to
post
The Go 6 might work? To be clear the Boox Android “improvements” are tolerable, not horrid; they don’t interfere with reading. They just also wouldn’t pass muster in a company with a decent UX team. If I go back to Boox, I’d definitely play with alternative launchers.
shop.boox.com/products/go6
comment in response to
post
(I have shattered two devices in a year after something like 10 years with my previous kindle? Bad luck.)
comment in response to
post
I liked the Boox hardware, and made the mediocre software work. When I dropped it, the Daylight had just come out so I took it as a chance to switch.
And the Daylight really is worth the higher price point, if the company weren’t apparently palling around with actively harmful people.
comment in response to
post
The previous paragraph in that newsletter was with an open antivaxxer podcast.
comment in response to
post
I found the Boox Android “enhancements” to be like most Android enhancements, actively detrimental to almost every part of the experience? Daylight’s were, sadly, the rare exception.
But yeah so far I suspect a Boox is the right answer.
comment in response to
post
From their most recent marketing newsletter, a paragraph that would almost be innocuous except for the hat and that this podcaster is a batshit conspiracy theorist
comment in response to
post
Make America Healthy Again, aka “we think America is unhealthy” (reasonable) “because too much vaccine and pasteurization and EPA and CDC” (and in Daylight’s case, too much EMF/blue light)
comment in response to
post
They didn’t take their vaccines
comment in response to
post
And MAGA is much more likely to treat big-city brown people like the Brits treated the Irish than like the Brits treated Bostonians.
comment in response to
post
It’s a pretty depressing book: we aren’t taught the history; the Founders had guts we are maybe lacking; and part of why Ireland remained trapped for another 150 years is because anti-Irish racism allowed much more brutal—and therefore effective— retaliation.
app.thestorygraph.com/books/41fe10...
comment in response to
post
Moderation is important, but concurrently it is inherent to massive centralized platforms that they replicate the power of government without the responsibility to the public.
(The 1st amendment (RIP, 1791-2025) anchored tech discourse in a privileged place.)
bsky.app/profile/rebe...
comment in response to
post
Ugh, they ruined a great phrase.
comment in response to
post
The fact that a hyphenated American isn't any less American than anybody else is literally the best thing about this stupid-ass country, the thing that most clearly makes it a redeemable project.
comment in response to
post
By mostly coincidence bought this yesterday flagsforgood.com/products/for...
comment in response to
post
This is California, California law, and California courts. So… yes, they do.
comment in response to
post
I live in the Mission, somehow missed the voter guide
comment in response to
post
Also why on earth is this a vote and who are these voters
comment in response to
post
(Holding off on more until the planned new bike, which will present new flag opportunities)
comment in response to
post
Have this in my bag already. Wish they made the California pride one in this size to complement my two SF flags.
flagsforgood.com/products/for...
comment in response to
post
Argh, I wish they let you search by size. Would happily fill an entire ebike flag pole with 12”x18”.
comment in response to
post
Generously, it’s b/c many of them haven’t been in the trenches at all (or at least not the trenches against CA-style fauxgressive NIMBYism).
Since they don’t recognize the book’s description of the problem, they assume the proposed solutions must serve a hidden, more nefarious agenda.
comment in response to
post
On the one hand, hooray if that actually happens; on the other hand, what a damning indictment of the CA Democratic Party.
comment in response to
post
xkcd compiling, but “walking between stops on a walking tour”
comment in response to
post
And then in those places that have made changes, do the work to run candidates in a way that takes advantage of it and shows the strength of the alternative system.
My biggest disappointment with RCV in SF is that people aren’t organizing slates or alternative parties.
comment in response to
post
I’ve been working with the legal team at Sonar to do some legal experiments with LLMs, and it’s been interesting and fruitful (more about that after my vacation). But I’m trying my damndest to also educate about the shortcomings at the same time.
comment in response to
post
Maybe DC has but SF basically hasn’t.
comment in response to
post
We put a lot of work into our metadata, such as having two rounds of manually validating the ToS of websites in Common Crawl, manually identifying trustworthy YouTube channels, and leveraging work by the BigCode Project and @softwareheritage.org to build the openly licensed subset of StackV2.
comment in response to
post
But if it’s the case, (1) that might be fine for you if your main use case is replication rather than broad reuse and (2) it’s certainly in part the fault of a legal community that hasn’t built helpful tools/process/documentation for this sort of data set.
comment in response to
post
My main thought in the original comment is that many data sets suffer from “we complied ourselves but have not done the work to make it easy for our downstream consumers to comply”, and I took Mike to be saying that this was the case here.
comment in response to
post
Please don’t read too much into it! Was responding to Mike while boarding a plane, so have not gone deep yet.
comment in response to
post
If we abandon the data field now, because no one of that sort can train *today*, then we also foreclose future training by those same groups.
comment in response to
post
IOW, I doubt no-legal-team groups will do meaningful training. But if low-barrier data sets exist now, then institutions with medium capacity and small legal teams (universities, medical centers, governments) will be able to do meaningful training soon.
comment in response to
post
I think part of the really fiendish problem here is that societally, we should want to be building low-barrier-to-entry data sets now so that middle-barrier-to-entry training can be done ASAP.
comment in response to
post
Sure, but fair use without a large legal budget isn’t a solution. So if @eff.org and @creativecommons.bsky.social want to establish a fund that pre-commits to defending data sets that meet certain pre-defined criteria (which could include due diligence on permissive licensing!) that’d be great…
comment in response to
post
Then the answer is… this is just impossible so you should sit on your hands while VC and F500 build the future? You should take on huge amounts of risk without even any appreciation? …?
I’m open to “attribution at this scale doesn’t make sense” but we have to offer something constructive instead.