best not find yourself saying "people have different definitions of opt-in"

[aud@fire.asta.lgbt]

@fasterandworse@hci.social "people have different definitions of opt-in"

maaaaaaaan even if that wasn't a bullshit cop out statement... they're still going against literally all of the "definitions"!

[aud@fire.asta.lgbt]

@fasterandworse@hci.social (I think a much more accurate way to phrase what people have is that people have differing opinions on how far away the LLM must be from the software they're using. They all agree that "opt-in" means what it fucking means, which is: "none of this happens unless I explicitly give my consent". They just differ on some slight particulars.

NOBODY thinks that means, "opt-in can mean opt-out").

[aud@fire.asta.lgbt]

@fasterandworse@hci.social I'm starting to get the impression that to advance in Mozilla requires not excellent project management skills, or managerial skills, or solid programming ability... but instead the ability to fucking gaslight the hell out of everyone.

Because I've met 'regular' people at Mozilla and they're great! But all their leads and C Suites are inevitably like this.

[davidgerard@circumstances.run]

@aud @fasterandworse now now, jake only talks like an abuser because he's being paid to

[aud@fire.asta.lgbt]

@davidgerard@circumstances.run @fasterandworse@hci.social "love your job, and you'll never work a day in your life"

[craignicol@glasgow.social]

@fasterandworse so the informed consent is as flexible as "true" in JavaScript?

I understand that people will have different tolerance for LLMs (naming tab groups feels less harmful than auto summarisation or browsing on my behalf), but that's absolutely not the message that I get from that quote.

[oggie@woof.group]

@fasterandworse
Does anyone want to take over/under on how long till you have to reaffirm that you still don't want ai?

They said explicitly that picking absolutely no ai will persist across updates, I am just curious how long till they show that was a lie.

Personal bet is 8 weeks, and then couched with 'well it's a major update so'. Then fro one global toggle which won't catch everything to ...

I mean, we had translation ( not perfect, but functional) well before LLMs. But not anymore!

[kentenmakto@mastodon.ie]

@fasterandworse https://www.youtube.com/watch?v=0_ipnwoPczQ

[womble@infosec.exchange]

@aud "I have altered the definition of 'opt-in'. Pray I do not alter it any further."

[thatdnaguy@genomic.social]

@fasterandworse people have different definitions of opt in. That's a good one. 😂

[angry_drunk@union.place]

@fasterandworse I work for a full-bore, Dominionist Christian, Trump-humping, AI-pickled freak…and I’ve seen less gas-lighting in my job than what comes out of that account.

[jargoggles@kolektiva.social]

@fasterandworse
So, like... is the Mozilla team saying, "Lots of people have different definitions of 'opt-in,' so we're going with the one that rapists use?"

[gbargoud@masto.nyc]

@fasterandworse @aud

I think the difference in definitions is "opt in to running the feature" vs "opt in to seeing the feature"

So a prominent button on every page that you have to scroll past is opt in by that definition because you still have to click on it to make anything happen even though its entirely in your face and makes everything worse

[brokar@mastodon.social]

@fasterandworse People only have one definition of opt-in. Yes, it's binary. Turn it on if i choose to or not.
The ones who have multiple levels of definition are corporations.
And this is only to rectify their personal agendas regarding data gathering.
But Mozilla is in good company with this statement, Microsoft for example operates like this for years regarding telemetry.

Also, the usage of "Blocked" instead of "Disabled" also speaks volumes, lets one assume it's enabled but just not shown.

[grumpybozo@toad.social]

@fasterandworse @davidgerard We in the anti-spam world have a name for people who make those sorts of arguments.

SPAMMERS

[davidgerard@circumstances.run]

@jargoggles @fasterandworse Jake Archibald is absolutely a guy you need to watch your drink around

but don't worry, he only does that because Mozilla pays him to

[davidgerard@circumstances.run]

@Oggie @fasterandworse also, they could have ... just not used an LLM for the translator?

[froztbyte@mastodon.social]

@Oggie @fasterandworse I’d guess longer, but only because I’ve seen how long features do end up in implemented but unsurfaced state (I’ve been watching some firefox dev around the profile switcher changes, and there’s things that showed up alongside)

as I’ve said elsewhere: moz has already made the choices on this months ago, and is now just working the plan to get them out

[anthony@indieweb.social]

@fasterandworse @HTeuMeuLeu @firefoxwebdevs I'm no GenAI fan, nor support Mozilla's leadership overall strategy. Yet I find Jake's explanation accurate: Firefox fans/critics have a different threshold of what is opt-in. Models included in the binary? Feature disabled but UI to suggest enabling it (as in https://mastodon.social/@firefoxwebdevs/116006613406440846)?

There's always a balance to find when introducing a new feature (AI or not):
- A mention in release notes, but most users will not discover it.

(continued)

[anthony@indieweb.social]

@fasterandworse @HTeuMeuLeu @firefoxwebdevs

- You can open a release notes modal/tab, but some complain that you're getting in the way.
- You can have some UI to promote it when the user is doing a related action, but some complain you're trying to force them to use it.

I feel that Firefox found a good balance introducing those GenAI features.

GenAI is very divisive, and sometimes we critics will be negative about a related subject without cause. I believe this is one of those situations.