yoasif

Immigrant native New Yorker. Product person. Into open data and open source. Ex-moderator of the Firefox sub-reddit.

Into #retrocomputing #firefox #haikuos #ux #design #openstreetmap #opensource #opendata #writing #retrogaming #linux #debian #btrfs #music

Blog: https://www.quippd.com
Photos: https://pixelfed.social/yoasif
My Blog on Mastodon: https://mastodon.social/@quippdblog
Donations: https://ko-fi.com/yoasif

Public Key

npub1zxttss5fffnmterjsuzcxmzty0watd6nggu33am9ycxtv7096u9qu2876u

NIP-05 Address

yoasif_at_mastodon.social@momostr.pink

Profile Code

nprofile1qqspr94cg2y55ea4u3egwpvrd39j8hw4kaf5ywgc7ajjvr9k08jawzspz3mhxue69uhhyetvv9ujuerpd46hxtnfduqs6amnwvaz7tmwdaejumr0ds772ejx

Publishing to

relay.damus.io

nos.lol

Author Public Key

npub1zxttss5fffnmterjsuzcxmzty0watd6nggu33am9ycxtv7096u9qu2876u

Show more details

Published at

2026-05-10T10:15:51Z

Kind type

0 Metadata

Event JSON

{ "id": "91681bc04c63cc93fdc37424ec0d3928c2ea29004c2158bf00e891b56d110872", "pubkey": "1196b842894a67b5e4728705836c4b23ddd5b753423918f765260cb679e5d70a", "created_at": 1778408151, "kind": 0, "tags": [ [ "t", "music" ], [ "t", "linux" ], [ "t", "opensource" ], [ "t", "firefox" ], [ "t", "ux" ], [ "t", "debian" ], [ "t", "opendata" ], [ "t", "design" ], [ "t", "writing" ], [ "t", "retrogaming" ], [ "t", "haikuos" ], [ "t", "openstreetmap" ], [ "t", "retrocomputing" ], [ "t", "btrfs" ], [ "proxy", "https://mastodon.social/users/yoasif", "activitypub" ], [ "L", "pink.momostr" ], [ "l", "pink.momostr.activitypub:https://mastodon.social/users/yoasif", "pink.momostr" ], [ "-" ] ], "content": "{\"name\":\"yoasif\",\"about\":\"Immigrant native New Yorker. Product person. Into open data and open source. Ex-moderator of the Firefox sub-reddit.\\n\\nInto #retrocomputing #firefox #haikuos #ux #design #openstreetmap #opensource #opendata #writing #retrogaming #linux #debian #btrfs #music\\n\\nBlog: https://www.quippd.com\\nPhotos: https://pixelfed.social/yoasif\\nMy Blog on Mastodon: https://mastodon.social/@quippdblog\\nDonations: https://ko-fi.com/yoasif\\n\",\"website\":\"https://mastodon.social/@yoasif\",\"picture\":\"https://files.mastodon.social/accounts/avatars/108/195/766/497/588/786/original/7727524e8abdfaf4.png\",\"banner\":\"https://files.mastodon.social/accounts/headers/108/195/766/497/588/786/original/c54fd7aeac85e31e.jpg\",\"nip05\":\"yoasif_at_mastodon.social@momostr.pink\"}", "sig": "0863733e9ac399dbfdfe147555a081c7ee41685f555523922318fa2417e15277fe346162361267c3907e9189096159bb8b907ea83c03c2cd2083db1973401288" }

Last Notes

2026-05-21 14:49:00 UTC

- reply

Glad to see it only took 5 years -- years after they deprecated it with hardly any analysis: https://www.quippd.com/writing/2021/03/21/supporting-compact-mode-in-firefox-proton.html
The removal of compact mode in Firefox back then is what led me to realize that some of the people building the browser didn't know what was in it.

2026-05-21 02:12:52 UTC

Disabled the #Cloudflare analytics on my blog https://quippd.com after reading this article: https://www.wsj.com/opinion/how-i-choose-which-cloudflare-employees-to-replace-with-ai-40a197e5?st=JGJffd
Thanks for telling us how little you care about the people who work for you.
#tech #layoffs

2026-05-07 17:18:39 UTC

- reply

The attachments show the result, not the process. People using the tools would be able to give us a better understanding of what is actually happening.

2026-05-07 17:11:14 UTC

- reply

Touché! I understood that "using" AI gave you the result - the topline here does the same "found and fixed".
We know the Firefox team didn't find the bugs themselves - that was AI - and the initial post implied (to me) an equivalency between finding and fixing.
I hoped to understand how much the AI had contributed to fixing the bugs, but it seems like we'll have to see another blog post for that.
Basically, what does "using AI" mean when fixing these bugs.

2026-05-07 16:26:13 UTC

- reply

You are the ones fixing the bugs, but your comment is pretty ambiguous -- your initial post says the bugs were fixed by AI, and then you say that you need a human author to write and review the patch.
Which is it - is it a human author or an AI author?
If it is a human author, can you really say that it was fixed by AI?

2026-05-07 16:20:37 UTC

- reply

You fixed the bugs with AI too?

2026-04-16 15:34:59 UTC

- reply

I wrote a little about #Mozilla pivot to #AI as well: https://www.quippd.com/writing/2026/02/09/firefoxs-ai-kill-switch-is-a-trap-how-mozilla-made-ai-your-problem.html
You may enjoy it.
I have largely moved to @npub1vpu…608s FWIW.

2026-04-09 01:19:50 UTC

- reply

I'm not sure that I love software patents, since they can lock people out of building interoperable alternatives to patented works. I think copyright is the right level of intellectual property protection for software - too bad it is being disrespected so widely (not just for software, obviously).

2026-03-26 01:49:54 UTC

- reply

Chromium browsers are obvious NOs...

2026-03-25 17:55:30 UTC

- reply

Sorry, I'm not going to waste my time trying to explain things to you when I'm avoiding working on an explainer post on the topic of LLM piracy and open source.
PS: I am not confused.

2026-03-25 17:49:42 UTC

- reply

LOL training open repos isn't piracy?
I'm clearly wasting my time with you.

2026-03-25 17:07:12 UTC

- reply

> I also don't understand how existing models become less useful. If the model solves a problem today, it will continue to do so tomorrow when a better model is available.
That is fair enough, but assumes that the problem domains will be the same in the future.
That is clearly what the model builders want, but to promote dependence, they need to cater to new use cases, not just existing ones.

2026-03-25 17:05:29 UTC

- reply

Sorry, how does this not continue to rely on piracy?
> We select approximately one million high-quality GitHub repositories based on criteria such as star counts, fork statistics, and contributor activity. Following Seed-Coder, we concatenate multiple source files within the same repository to form training samples at the project-level code structure, preventing the model from learning only isolated code snippets and promoting understanding of real-world engineering context.
https://files.mastodon.social/media_attachments/files/116/290/890/442/039/919/original/c3b74d22d56daa1c.png

2026-03-25 15:39:37 UTC

- reply

Training will always be a bottleneck, and ongoing piracy from China isn't guaranteed, especially as they pose a risk to administration affiliated businesses.
See the ban on foreign routers from yesterday: https://www.usatoday.com/story/tech/news/2026/03/24/fcc-bans-new-router-imports/89300646007/

2026-03-25 15:37:51 UTC

- reply

That seems doubtful, with Jensen Huang claiming we've "reached AGI" - the existing models are going to pay off before we see even more investment into something that is just a glint in the eye of most engineers working in this field - it ain't happening without MUCH more investment than we're already getting.
For the existing models, they tend to become less useful as the "state of the art" changes, so ongoing piracy is required (see the deals with Wikipedia from big tech).

2026-03-25 13:41:14 UTC

- reply

That or they believe that you won't have the same capacity or ability to pirate the data required for the models to be useful.
They get a carve out, you get a parrot that still needs training.

2026-03-13 19:02:24 UTC

- reply

Mozilla claims that you can use a local LLM: https://www.reddit.com/r/firefox/comments/1rri5e7/i_tried_firefoxs_new_ai_smart_window_in_a_beta/oa0q3ha/
But I found a forced login to Mozilla/Apple/Google accounts when I tried it.
https://files.mastodon.social/media_attachments/files/116/223/406/986/336/868/original/9f8bc10d0fbd1b0e.png

2026-03-05 20:26:12 UTC

With the conversations happening around clean-room implementations of #opensource code re-licensed as something different in #python chardet - you may want to read about how this is happening across the entire ecosystem, and how the glorious big tech companies are bankrupting open source:
https://www.quippd.com/writing/2025/12/17/AIs-unpaid-debt-how-llm-scrapers-destroy-the-social-contract-of-open-source.html
For context on chardet: https://github.com/chardet/chardet/issues/327
Boosts appreciated!

2026-02-26 17:37:01 UTC

#Mozilla released #Firefox 148 featuring the #AI "Kill Switch" that has been making the rounds.
The design of the switch is meant to absolve Mozilla of any responsibility for your use of its AI.
It is a trap.
It is meant to switch discussion from whether Mozilla should be doing AI, to whether you can *choose* to not use it.
Don't believe their lies.
This is a long post, and if you click through, I recorded a video: https://www.quippd.com/writing/2026/02/09/firefoxs-ai-kill-switch-is-a-trap-how-mozilla-made-ai-your-problem.html
Hope you enjoy and please boost if you enjoyed it!

2026-01-14 20:09:58 UTC

Simon Willison on porting OSS code:
> I think that if “they might train on my code" is enough to drive you away from open source, your open source values are distinct enough from mine that I’m not ready to invest significantly in keeping you. I’ll put that effort into welcoming the newcomers instead.
https://simonwillison.net/2026/Jan/11/answers/
This feels very much like colonialism; take over all the #OSS code, drive the original developers away, and give the colonizers the code as a welcome present.
#AI #llm

2026-01-07 01:21:13 UTC

- reply

Wegmans doing retina scans already marks them as bad in my book, and I went to Wegmans a ton when I lived in Buffalo (for school).

2025-12-17 16:49:36 UTC

Which #Firefox fork are you moving to as #Mozilla gambles their userbase in search of people who will train big tech #AI for free?
Boost for reach!

2025-12-17 01:14:09 UTC

Is #Mozilla new CEO actually threatening the removal of MV2 based ad blockers in #Firefox on his FIRST DAY as CEO?
"He says he could begin to block ad blockers in Firefox and estimates that’d bring in another $150 million, but he doesn’t want to do that. It feels off-mission."
Why is this even on your mind?
🤯
https://www.theverge.com/tech/845216/mozilla-ceo-anthony-enzor-demeo

2025-12-13 00:06:49 UTC

- reply

I was reading https://decrypt.co/351694/us-tech-giants-unite-battle-chinas-open-source-ai-dominance last night and this caught my eye:
"The foundation addresses a strategic vulnerability: Chinese open-source models create dependency—developers worldwide build on that infrastructure, reducing reliance on U.S. cloud providers and APIs."
So clearly what this means is that "open" doesn't mean open source, because Chinese models are open source -- it means "wedded to American big tech".
Truly face-palm worthy.

2025-12-09 19:53:47 UTC

- reply

Open source is being damaged left and right: https://www.quippd.com/writing/2025/12/08/mozillas-betrayal-of-open-source-googles-gemini-ai-is-overwriting-volunteer-work-on-support-mozilla.html

2025-05-11 17:25:47 UTC

- reply

A theme of my latest post: https://www.quippd.com/writing/2025/05/07/firefox-forever.html

2025-04-14 17:40:12 UTC

Looking to help my dad out with some gardening and want to borrow a post driver. Does anyone have one I could borrow for a day? #gardening #nyc #queens

2025-03-06 16:41:14 UTC

- reply

and yet you won't explain why the differences between the first version and the second exist.
Why wasn't the language the first time around scoped to "what they have been doing"?
Why did they claim that the original text was necessary to support their business?
If no mistake was made, why was there a change?
All clearly cannot be as before -- even between the two versions of the terms!

2025-03-06 16:24:58 UTC

- reply

Sincerely, what happened, then?

2025-03-06 16:20:18 UTC

- reply

I don't know how you have the clarity over Mozilla's ad practices to know that they aren't selling data when they can't even muster that language.
I understand not wanting to rush to judgment, but Mozilla came right out of the gate with an extremely broad license grant that they defended as required for their business before backing off within two days.
We already know Mozilla tried to get more legal privileges than required!

2025-03-06 14:50:40 UTC

- reply

If that is the case, why did they change the IP grant within TWO days?
https://www.quippd.com/writing/2025/03/05/the-mozilla-firefox-new-terms-of-use-disaster-what-actually-happened.html#the-policy-change
If the language on Wednesday was already tailored to what they are already doing, why is the update provided on Friday less broad?
Did they make a mistake on Wednesday or is it wrong today?

2025-03-06 14:38:22 UTC

- reply

PS - you say it is "much better to be broad" - that clearly depends on your perspective - if you are Mozilla, you would like for users to give up their lives - why not, it is better for them.
For Firefox users, I don't see what it gets THEM to be yoked to a broad agreement. I don't see how it is better for them.
Given that these are rights that the users are granting to Mozilla, I don't know that users should automatically believe that a broad agreement for usage of their data is good.

2025-03-06 14:35:29 UTC

- reply

I can't even tell what is happening, yet you are somehow skipping ahead to the legalese not covering what is already happening.
I asked you specifically if Mozilla needed to make the ToU change that they already made -- if they made a mistake, won't they need to update it again with the old language?
If the new language is correct, why was the old language so over-broad?
It seems to me that the legalese means things, and Mozilla changing their language is acknowledgement of that.

2025-03-06 14:18:23 UTC

- reply

The legalese opens users to new abuses. That is new, is it not?
I agree that "nothing has changed", yet everything has - previously, Firefox users weren't agreeing to have their data sold (that is what the legalese says), and product updates can make that possible.
Technologists are annoyed because they know how technology can be abused to expand abuses to fit within the letter of the law.
I'm curious - would you say that the ToU update that has already been pushed was unnecessary?

2025-03-06 14:10:10 UTC

- reply

Frankly, their posts and updates are clear as mud -- can you identify what data is being transferred? I can't.
All I can tell for sure is that Mozilla is linking personal identifiers, and I don't know why they are claiming something in their blog posts that their lawyers won't support in their ToU.
See this post which surprised me: https://www.quippd.com/social/posts/2025/03/02/who-else-is-surprised-that-mozilla-uses-technical-and-interaction-data-for-advertising-purposes-on-the-new-tab-page.html
If you can show me how our data is used, that would do a lot to clarify things.

2025-03-06 13:50:15 UTC

- reply

How do you know the first? It is not at all clear that Mozilla is even saying that.

2025-02-01 19:36:15 UTC

- reply

Have you done any troubleshooting steps? Considering I'm not seeing widespread reports of this, it feels more like an issue on your setup.

2025-02-01 16:04:46 UTC

- reply

Unfortunately, I have no need for YouTube Studio or Steamworks. People seem to be able to use Bluesky - at least I haven't seen any widespread reports about it, and I feel like that would have come up: https://support.mozilla.org/kn/questions/1474705
Are the sites reporting that they don't support Firefox, or are you experiencing something different?

2025-02-01 05:03:24 UTC

- reply

What site? You can always report to https://webcompat.com

2025-01-28 15:18:02 UTC

- reply

Trash guide.
> (They don't read bug reports anyway.)
Uh, they do. Ask me how I know.

2024-12-19 02:20:40 UTC

- reply

Do you think this applies the same way to #reddit and its API change?

2024-10-22 20:30:55 UTC

- reply

Amusing that you can't just admit that Vivaldi shows more ads no matter what you do to it, all while calling other browsers "adware" (even though Vivaldi ships ads).
You could try being honest. I don't bother trying to defend things I don't like - it makes life simpler for me.

2024-10-22 20:22:17 UTC

- reply

It is a lie to say that Vivaldi + whatever you add to it results in a browsing experience with fewer ads than Firefox + uBlock Origin.
I don't have any problem calling that out, no.

2024-10-22 20:14:43 UTC

- reply

How is your blocking improved when the Vivaldi blocker is worse than uBlock Origin? More lies.
In any case, I have no issue running Firefox and expecting a good browsing experience. I absolutely expect that, and that is why I use it. That is also why I don't use DNS based blockers - those often give me a worse browsing experience.

2024-10-22 20:08:08 UTC

- reply

You typed a lot of stuff to confirm that Vivaldi shows more ads today than what is possible in Firefox. Adding a DNS based blocker? What kind of joke is that?
That's the kind of stuff you recommend when you don't have access to a good ad blocker.
C'mon.

2024-10-22 19:58:25 UTC

- reply

Vivaldi ships a bad ad blocker and Firefox has the full featured uBlock Origin. It is also open source, so it is hard for me to understand why the standard is lower here for the amount of adware (there are distributions that release reconfigured versions of Firefox without sponsorship -- not possible with Vivaldi).
Vivaldi has more ads TODAY than a Firefox with uBlock Origin. That is inescapable - not only that, I doubt we EVER see a blocker with the power of uBlock Origin on Vivaldi.

2024-10-22 19:41:10 UTC

- reply

Vivaldi has ads. Based on your definition, it is adware. Why lie?

2024-09-25 19:32:24 UTC

- reply

Are you sure about that?
https://old.reddit.com/r/browsers/comments/1fefqyq/ive_tested_21_browsers_multiple_times_in/
PS: I'm not even a believer in benchmarks for browsers in general because no one can know beforehand what pages you will load, but...

2024-08-08 20:19:46 UTC

- reply

Glad it is working for you!

2024-08-08 13:35:17 UTC

- reply

You should be aware that you are signing up for an experience in masochism without any perceived marginal benefit.