Hacker Newsnew | past | comments | ask | show | jobs | submit | maxloh's commentslogin

I thought Q4_K_M is the standard. Why did you choose the 6-bit variant? Does it generate better input?

There is no standard.

The higher quantization - the better results, but more memory is needed. Q8 is the best.


FP32 is best, although I wonder if there isn’t something better I don’t know about. Q8 is for the most part equal to FP16 in practical terms by being smart about what is quantized, but iirc always slower than FP16 and FP8.

Some Opus models were free on Copilot, and in my country you cannot attach a repo to Gemini, that is limited to their premium offerings.

Which Opus models were free on Copilot?

Haiku, right?

VLC is GPL licensed and open source. Maybe that's enough since you could compile it yourself anyway?

The adage was "If you're not paying for the product, you are the product." Now enterprises are paying to become the product. That's ridiculous.

Nope. You can never verify they run the same code from their repo. You cannot physically access their system after all.

Illegal actions are often hard to prove, and yet laws somehow work in general. Same here with obeying the license.

In contrast, the Multi-Account Containers system is the primary reason I avoid Firefox.

While it is meant to be an alternative to Chrome's profile switching, it is more a workaround than a complete replacement. I need entirely different sets of extensions for personal, work, and school environments, something containers can't do.

Firefox's actual profile support is beyond terrible. To launch a separate instance, Firefox requires many more clicks than Chrome, all within a Windows-2000-style UI. Not to mention that there are weird glitches in their implementation.

Firefox is not usable for me until they actually spend time improving their multiple profile support.


I definitely have not had that experience, although use FF for personal, various work, and various educational places.

None of those have required me to install a particular extension..

Of course thats not to deny your experience!

The only time profiles ever come into it, for me, is using web driver, playwright, or whatever.

I guess maybe the usage stats dont support making the profile selector better.

But also, maybe its a thing they would accept a change for?


This is not meant to be an alternative for Chrome's profile switching. It's a different use case entirely.

As you yourself mention, Firefox has actual profile support, which may not be as good as Chrome's, but at least compare like for like.


Myself the profile support is the absolute worst thing about Chrome. I just want to log into some web site, I don't want to fight with the profiles to get things done.

For those few applications where I really would need profiles I will just open a different browser, so I still keep Edge/Chrome/Opera around for that rare situation. I don't need something that complicates my life every single click but it is the whole ideology of the Google Economy that they want you to spend 1% of attention on things that matter to you and 99% on things that don't.


Firefox has a new Chrome-like profiles support as of v149.

https://support.mozilla.org/en-US/kb/profile-management


LMFAO. Containers are not for profiles-purpose. Everyone who needs profiles know this.

And Firefox now needs 2 click to switch profiles.


Mozilla is doing exactly what you’re describing. They need revenue to ditch their direct financial ties to Google (and I wonder if they hire those high-salary executives solely in the hope of generating that revenue).

These AI products, along with all previous failed attempts, are just them trying to gain enough revenue to remove that dependency on Google.


And you your point, AI is probably eating search and with it the prospect of search licensing revenue. Not sure yet what paradigms will be most important to the browser experience but it's critical to get in early and make the inevitable early mistakes and work through them.

I truly hope that is the case. I feel they're going about it from all the wrong angles, but I sincerely hope it works out in their favor. Their funding model and governance seem inherently broken and have for a long time...

Mozilla needs money to support the development of Firefox (and the payroll of its high-salary executives).

For now, they mainly rely on Google for that money. Google pays them to avoid antitrust cases, to show the courts that they are not a monopoly and that "alternatives" exist. For example, the DOJ once proposed that Google be forced to sell off Chrome.

However, if another entity has control over your budget, they also have control over your product. If Firefox becomes "too good" to be a true competitor in the consumer space, the funding might be reduced or even cut off.

Creating a new source of revenue allows Mozilla to improve Firefox even beyond the point Google feels "comfortable" with.


Mozilla could stop doing everything else and slow burn their existing $1B into developer salaries over the next decade. They are actively choosing not to.

And then what? Just go bankrupt after a decade? That's entirely unsustainable.

1. It's unfair to assume that their primary funding source stops in one scenario and not in the other.

2. 1 billion dollars is a lot of money. Even the interest off it is huge.

3. 10 years is a very long time in tech.

4. I would greatly prefer the money Mozilla earned due to Firefox being a thing was put into developing Firefox, yes. The current Mozilla organization seems to be a mechanism for providing third homes for the executives, starting projects nobody wanted them to start, sullying the Firefox brand with them, and then abandoning them. There's a VC cancer infesting the supposed "free software community" called Mozilla.


>1. It's unfair to assume that their primary funding source stops in one scenario and not in the other.

Wait, what? I thought your whole premise from one comment ago was that they "stop doing everything" and exclusively slow burn away their endowment. They're dead by 2029 if they do that.

If they don't do that, then you're just talking about how they currently operate.


The money paid by Google so Chrome does not look like a monopoly is earned by Firefox and specifically for Firefox to exists as a viable-enough competitor. If anything, maintaining Firefox properly is the branch that earns that money.

Mozilla should stop doing all these side quests -- look at their track record! -- and they should get rid of the fat executive layer. They should transparently report what they're using their money, instead of saying they burn hundreds of millions of dollars in "software development" while firing the Servo developers.


Actually, it makes much more sense if Google pays Mozilla to maintain an alternative that never becomes truly competitive.

I don't think actual competition benefits Google in a commercial sense. If we considered the situation purely rationally, Google's most logical decision would be to use their budget as leverage, threatening off the books to prevent any strategy that might make Firefox "viable again", under the assumption that Google focused primarily on market share, while Mozilla focused purely on survival.


Yes, Google definitely has an incentive to keep Firefox inferior. If it became a real competition, that money flow would likely stop.

However, if Firefox drops to ~0% usage, it will definitely stop, as that ruins Google's monopoly defense, which is the motivation for it! Firefox usage is supposedly already as low as 2.33%.


Even worse, they would go bankrupt after like 2-3 years. And almost no serious non-profit has an endowment that can fund them in perpetuity. They are almost always firewalls that buy time/safety in response to crises (e.g. financial crisis or covid), meant to run in parallel with ongoing revenue that comes in regularly.

It's great if you can get your endowment so big you never have to worry about revenue but outside of, say, elite universities or middle eastern sovereign wealth funds that rarely happens.


I have the same experience with my RX 5700. The supported ROCm version is too old to get Ollama running.

Vulkan backend of Ollama works fine for me, but it took one year or two for them to officially support it.


Note that this is fundamentally different from the Astral acquisition. At the end of their announcement, they stated:

> Cirrus CI will shut down effective Monday, June 1, 2026.

And earlier in the article:

> Joining OpenAI allows us to extend the mission we started with Cirrus Labs: building new kinds of tooling and environments that make engineers more effective, for both human engineers and agentic engineers.

It isn't a product-led acquisition, but more a talent one.


This is kind-of neat too, at least in the near term:

> In the coming weeks, we will relicense all of our source-available tools, including Tart, Vetu and Orchard under a more permissive license. We have also stopped charging licensing fees for them.


That part is amazing. Back when I first heard of tart I thought it was amazing, with the one downside being the license.

Hopefully development on it continues, or a community maintained version keeps it going.


This is making my current work 100 times easier. Very welcome timing.

This is awesome because I flippin love tart.

This is a huge deal! I secretly hoped for this. :)

Just want to note that we will continue maintaining and improving our virtualization solutions actually with even greater attention. SaaS options like Cirrus CI and Cirrus Runners will eventually wind down so we can focus on incorporating pieces internally.


You’ll be pleasantly surprised. Updates in the coming weeks.

> In the coming weeks, we will relicense all of our source-available tools, including Tart, Vetu and Orchard under a more permissive license. We have also stopped charging licensing fees for them.


MIT or Apache2 or FreeBSD licenses would be preferable in my case, but GPLv2 or even AGPLv3 (if you have to) would work.

If you’re taking requests…


That sounds like a British phrase for pimping.

If your scope includes making the Codex web app environments have additional functionality I look forward to it. More enterprise features and yaml backed pipelines.

If you are interested in yaml backed pipelines check out this open source tool I built for exactly this purpose:

https://github.com/smogili1/circuit


For now.

I mostly clicked the link because I was curious if Cirrus Labs operates Cirrus CI and if so how that would be impacted.

Looks like I’ll need to move the FreeBSD CI jobs for open source projects I maintain to another solution. Anyone have suggestions for alternatives?


I guess qemu over Ubuntu-latest from GitHub Actions running freebsd, but it will be a bit flaky

Same for me. I liked the native BSD runners, qemu on gh is much slower.

It could also be a suite of product acquisition, the CI could be a product OpenAI is interested in having, but not sell.

Yeah. Much like Astral - acquiring both the product (because they need to use it internally, but don't care about trying to resell / market), and they also want the talent to keep maintaining it / add features they want.

>It isn't a product-led acquisition, but more a talent one.

I am pretty sure OAI mostly cares about their virtualization IP for MacOS. They already extensively use WSL2 for sandboxing Codex on Windows, and I imagine they want something similar for Codex on Mac.


Is Sam or family an investor in them anyways?

We were 100% bootstrapped with no outside capital or support/advisory.

Congrats, I am sure you and team made some $$$.

Consider applying for YC's Summer 2026 batch! Applications are open till May 4

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: