craigmart

I believe that comparing LLMs with traditional deterministic software is fundamentally misleading. It is extremely difficult to truly interpret what LLMs do internally, and as of now, nobody fully understands it. Even if you trained the LLM yourself, there is no source code you can simply read and learn from.

Sure, having information about how these models were trained is helpful for reproducibility, but it is basically impossible for anyone without substantial capital and access to the same (likely copyrighted) data to reproduce the model. For normal users, owning the model weights essentially means owning 100% of the model, you can inspect and study the weights in much the same way as the lab that produced the model can, you can modify the weights, and you can use and distribute them if the license allows you to

It's up now

I strongly agree with antirez. I believe that an intellectually honest programmer should recognize where their knowledge and experience is mostly beneficial, and at this point (and we've seen clear signs for some time) coding is largely solved. Which doesn't mean that you can just prompt at a very high level, but that iterating with the LLM over decisions, designs and tests, will essentially allow you to not write a single line of code by hand. Software engineering is not just coding, and it involves a whole set of other tasks requiring direction and creativity that can greatly influence the quality and impact of the software.

The most common arguments against this view seem to arise either from ideological resistance, which I understand given how painful it can be to see one’s job at risk or an important part of one’s identity taken away, or from generalizing a small number of experiences with LLMs to the technology as a whole. In the latter case, those experiences may also be heavily conditioned by the user’s inexperience in working with LLMs, or by the specific use case in which they were applied. There are still certain tasks that not all models can handle reliably as of now, however some can (usually the most expensive), and they will likely continue to improve over time.

This is something we have known for a very long time, and companies are not trying to hide that either. They do it to avoid letting competitors train their models on the CoTs

I had checked that page not long ago, and as far as I remember there were many "red" or "orange" days in the past 3 months. Now it's all green. That's concerning

The "Stop. You know nothing." comment was apparently a reference to this tweet. https://x.com/CodyRhodes/status/980680154098757632

Embarassing nonetheless

You're right, pardon my cynical remark. I'm just disillusioned by the promises of most tech companies

I had checked as soon as I found out about the news the other day and it was there. I just checked on wayback machine and you're right, it was removed for some time. However, if they're willing to put back that claim immediately, I doubt that their intention was to drop the free plan anytime soon, but probably it was to incentivize people to use the paid plans. Enshittification must happen sooner or later afterall, but fortunately vaultwarden exists and the export feature is highly unlikely gonna be removed immediately as the free plan disappears, so people could just switch to a third-party or self-hosted backend as soon as that happens.

This is an incredible overraction over a minor change that did not even happen. You can still find "Always free" in the pricing line of the very same page everyone keeps linking as proof https://bitwarden.com/products/personal/#whats-the-differenc...

Edit: it actually disappeared for some time but they put it back on May 18

snapshot from May 15: https://web.archive.org/web/20260515190646/https://bitwarden...

snapshot from May 18: https://web.archive.org/web/20260518183728/https://bitwarden...

they haven't backed down, you find the "Always free" claim in the very same webpage OP linked https://bitwarden.com/products/personal/#whats-the-differenc...

Except for guessing the right continent (not that remarkable), mine is so majestically wrong that I would either dislike or fully hate all of the products I got recommended.

comedy is a completely different thing than natural tone. I agree that they’re incapable of coming up with decent jokes

You can make an LLM sound very natural if you simply ask for it and provide enough text in the tone you’d like it to reproduce. Otherwise, it’s obvious that an LLM with no additional context will try to stick to the tone the company aligned it to produce

Schools?

I totally agree. The authors have found a security vulnerability in a mechanism that wasn’t even trying to protect the software from attacks. I don’t see any relevance in this paper except being a good exercise

The point is that you don't have to trust them because the client (where the relevant cryptography is performed) is open source and the fact that my links point to signal.org is completely irrelevant, those blog posts are just ways to advertise facts that are freely verifiable. You can read the source code to check the implementation of sealed senders or how the social graph is handled.

NSA can hack into Signal's infrastructure, and what they will be able to gather are the same information provided by Signal in reply to subpoenas (the whole list here https://signal.org/bigbrother/), because everything else is end-to-end encrypted.

You are speaking of metadata as if all metadata is equal. Signal does collect phone numbers (even though, since usernames have been introduced [1], this can be made opt in from now on), but not the contacts or social graph, neither many other relevant metadata [2]. What they can gather from this, is only when the specified phone number registered to signal services and its last connection to the server [3].

So, if you can call "metadata exchanging app" an app that simply has a list of numbers registered to the service, without any metadata assigned to them except their last access, the same label could be assigned to a much larger number of services.

It may not be anonymous, but it can hardly be disregarded as private.

[1] https://signal.org/blog/phone-number-privacy-usernames/

[2] https://signal.org/blog/sealed-sender/

[3] https://signal.org/bigbrother/central-california-grand-jury/

one side of the "debate" seem to be taken solely by Joel Moskowitz, who's cited in every article that promotes the message of "phone radiation is harmful"

looks real to me

Could you expand on why people shouldn't compile their kernel? I think it's fairly useful to compile their own to get a better understanding of what the kernel does and to better suit everyone's needs. For example, if I have little free space on my boot partition and I have my disk encrypted, I want my kernel to be as small as possibile, so I will deselect every driver I don't need. Or maybe the driver for my new device is not included in the kernel builds of my distribution.

Not only I woundn't say that most people shouldn't compile their kernel, I would say that most linux users* should do it at least once, so they can understand the power they have compared to closed-source operaring systems.

*with linux users I mean users that use linux as their main operating system, not people that do ssh once in a while or rarely boots their linux partition

What kind of protection does Facebook Container have other than deleting cookies outside of the container?

For my case Total Cookie Protection is enough, but if you want the same protection of Facebook Container for every website (i.e. session cookies which are deleted each time you restart the browser) you can install Cookie AutoDelete or use the built-in option to delete cookies at restart (whitelisting websites where you need permanent cookies).

Although I don't agree with the article, I can understand how someone could start hating work and the urge to be always productive. Our economy promotes competitiveness, which of course creates progress and brought us where we are now, but I think covid worked as catalyst for all the pressure the workers were feeling because of this. And when you are so distressed by the system it's way easier to believe that you don't have to work as hard to be happy instead of starting to push harder

Sorry, I meant software in general but wrote "desktop applications" instead. Anyway the sentence is still valid, even if you'll have to implement other things such as the graphical interface, the POSIX compliant code won't need modification

>why in the heck is he going for POSIX compatibility

Because existing desktop applications can be ported to ToaruOS

>why not a safer microkernel, keeping everything in userspace?

This is a design choice, microkernels aren't necessarily better than hybrid, they're slower, harder to debug and process management can be complicated

So this service consists basically in paying to compromise E2EE, adding an attack vector, storing my otherwise encrypted conversations on centralised servers, using an arguably worse client with less features just for the convenience of not having to open 3 separate apps? Who is this for?

Donated to Signal and I intend to donate again to other projects as soon as I can. Unfortunately I'm not working so it's not that easy, but I'm doing my part to report bugs for every software I use, even using beta versions.