Mistral is pretty much toast? Their models perform poorly and I'm not sure why anyone would use them. Maybe there is a catching up point somewhere in the future, hopefully.
7,857 karma · joined January 6, 2011
If we are going to go down the path of mandating legal liability on software makers of a neutral communication medium, then the EU should just break the commercial web.
There are complaints on this story, and on the recent one about the fsf phone project about how inconvenient it is to not be able to access banking apps on your mobile phone. I can't be bothered to enter my banking password every 30 minutes on my desktop! What, I'm supposed to have two phones?
The first thing someone is going to do when they steal your phone (after they saw you enter your password in public) is open your banking and money apps and exfiltrate as much as they can from your accounts. This happens every single day. None of those apps should be installed or logged in on your phone. Same goes for 2FA apps. That's like traveling with Louis Vuitton luggage which is basically a "steal me" sign.
That's the most basic stuff for people who aren't a CEO of a company that is in the crosshairs of state sponsored espionage attacks.
The problems with "bare bones secure OS" device remain the same from a physical access standpoint: social engineering, someone sees your password, steals the device. But otherwise, yes, the devices you install a bunch of spyware/adware games on and take to bars should not be the ones you are doing your banking, 2FA, work, etc on ever.
The fundamental problem with recommendation engines are the platforms are forcing content on the users based on what increases engagement (and possibly ad $) on their platform -- not what is valuable to the user.
If I'm following a particular blogger, and their signal to noise ratio is 9:1, if they recommend an article it would be vrey likely I am also interested in it. If I'm following a larger group of people, I may be interested, in aggregate, what they are interested in. All very basic stuff that seems to have been abused and forgotten. Other than e-mail, rss is the odd man out in 2025 which also makes it so much more valuable.
Any third party RSS feed service could ingest these feeds and spit out whatever other type of recommended feed they want. The could just view a hn/reddit style feed solely with counts of the people they want to follow along with "vote" counts and some decay algorithm. They could have a chronological feed that just shows everything. The users are in control again, not Meta/Facebook/IG, Reddit, Bytedance, or etc.
First, the newsboat DB is SQLlite so it's easy to access. I wrote a few scripts that built static HTML pages of all of the feeds along with a feed style page. I copied the HN page html/CSS, which to me feels like the maximum compact view while still being readable.
Now I have a bash script that will refresh all of the newsboat feeds and then open the static HTML page in the default web browser.
Thinking a lot about the impact of "vibe" coding recently made me wonder why anyone needs to be locked in to not just a set UI but any sort of rigid external control of how the user sees and interacts with information.
I want to see all my news chronologically. Now in a single place I have hn, lobste.rs, numerous twitter/x feeds, mastodon, and a lot of other blogs all visible chronologically. If someone was noisy I could apply keyword filters to the feeds and block certain things out. If I wanted to, I could put this on a server and access it through my phone.
This was definitely "the easy" way to do this. You could raise $1 million and do the same thing the hard way.
Microsoft is setting fire to the bridges and those will be users which they will never be able to get back again.
Linux works great for gaming except some anti-cheat stuff which probably won't be legal anymore anyways in Europe under the PLD.
As others have pointed out, for iOS Monal is good.
The email spam issue is trivially solvable with a contact whitelist, which is a UI issue. Email as it is right now is definitely very usable, but keep me the hell away from anything from Microsoft.
IRC is alive and small. On the optimistic side it outlived Skype. Maybe 25 years from now IRC will still be working and Discord will be dead. There has been a lot of buy in on Matrix, but I'm unconvinced the protocol is going to thrive long term due to design choices made.
Reddit is doing what Usenet did. In my ideal world, reddit would be part of the fediverse along with Usenet & Twitter and the UI would close to hn.
When its an algorithm, the user is incentivized to produce content in order to increase their chances of getting a hit. Secondarily the loss of visibility increases value of advertising on the platform. It's a lose-lose for users, first they are forced to use the platform more for fear of missing something, second the user has to post more to get any reach. The platform wins on increased engagement, overall content depth, ad revenue, and the ability to sneak in a whole lot of shit the user never was interested in or followed. Facebook & Instagram now are functionally high powered spam engines.
Interestingly the FT has an article today about a drop in social media usage ( https://www.ft.com/content/a0724dd9-0346-4df3-80f5-d6572c93a... ) - one chart titled "Social media has become less social" shows a 40% decline since 2014 of people using social media to "share my opinion" and "to keep up with my friends" In many ways, what is being referred to as social media has become anti-social media.
An algorithmic email feed would be useless, as would any sort of instant messenger, yet that's exactly what social media turned in to. Twitter/X is teetering in that direction. The chronological feed still works and is great. Anyone who posts a lot and doesn't balance out the noise with signal I just unfollow.
I might be wrong about this one, but one outcome of generative AI might be an engagement cliff. Some users will be very susceptible to viewing fake photos and videos for hours (the ones still heavily using FB likely are), but others may just choose to mentally disengage and view everything they see on FB, IG, Tiktok as fake.
User mentions they didn't sleep well. Model delivers jarring information right before user's bed time. Model subtly suggests other sleep disruptive activities, user receives coupons for free coffee. User converts for ad for sleeping medication.
(This is already happening, intentionally or not)
Notably the open source models OpenAI released right before gpt5 are likely good enough to be substitutes for 95% of typical ChatGPT use cases.
> How much longer will free and open source be tolerated?
I don't think they have a choice. Imagine what would happen to Google if half their software stack was Oracle and the EU had backdoors in to all of the management and CEO's devices and private communication. Why not use Chat Control to verify that they are complying with the spirit of EU law? Turn on the remote microphones while they are at it too.
On one hand we can lament the death of open source. Yet, open source has never been healthier. There has never been more open source software available to use and in development. Even when in it comes to AI, the best open source models are actually really damn good, better than anything that existed roughly 12 months ago. As much as Google, Apple, and Microsoft want to force you in to their closed ecosystems they fear being locked in to their competitor's closed ecosystems even more!
This could be a 10 page comment, but yes, the regulatory environment is a real threat to open source and the open internet in general. Most of those threats have been coming from the EU, with things like Chat Control and PLD. Which is unfortunate, because the future of the free world will rest entirely with the United States (Also possible that the EU will be dissolved, the monetary union will have a very difficult time during the next financial crisis.)
On the other hand, software developers and users, have become too reliant on Android which is functionally a fake open source project now. I can't think of a stronger incentive to stop Android development than telling them you can't develop here without paying us.
For my personal projects I have a list of difficult bugs I keep that LLMs can't solve. Right now that list is empty. Anyone using LLMs for coding, using the best tools and practices, can see what a massive capability leap has occurred in the past 9 months.
If the US is going to end up "losing", it is going to be first through power generation -- China's coal plants produce more power than the US as a whole. Robotics is a whole other topic and shouldn't be bundled with LLMs.
One thing I've grown concerned about, after watching the Twitter migration fizzle out, is we can imitate the old internet on a small scale, but on a large scale it just doesn't work. For Twitter specifically, the outcome was even worse, many users just migrated to other more centralized services or existing monopolies (like Instagram.)
Users are too used to being able to instantly stream 4k HDR 60fps. They are too used to limited amounts of spam. They are too used to having most non-agreeable content filtered. All of this stuff that big tech delivered now is replicate-able at the cost of tens of billions of dollars. The only business model that can pay for that is owning a giant ad platform.
Thinking about all of the issues the EU has had enforcing things like GDPR, which big tech companies largely haven't followed for years or straight up lied to their customers about, along with a possible failure of the DMA now due to tariffs.. and yet on the other side of the Atlantic, the US utterly failed to ban or control Tiktok. Endless announcements of upcoming deals that were either lies (Oracle protecting American's data) or postponements.
Meanwhile, all of the spam, hacking, bots, and DDoS attacks persist and grow, along with layer upon layer of (probably intentionally) poorly written and often conflicting legislation across multiple jurisdictions have truly made it impossible for the internet as it was designed and meant to exist to continue. (Sure you can just set up a basic web forum like you could do 20 years ago, not use Cloudflare, not host it at a major datacenter, and ignore all of the GDPR and age verification laws, but good luck. Hell, it doesn't even sound like it's really legal to run a Mastodon server anymore.)
One small hope is that if internet companies follow any pattern we've seen in other industries, when the growth ends, the managers will switch to tearing the conglomerates apart in to pieces and selling them off. One day CloudFlare might be split in to 30 pieces, along with Alphabet, Meta, and Amazon. But it could be a while.
The exponential growth of compute and data continues..
As a side note, if anyone I'm communicating with - personally or in business - sends responses that sound like they were written by ChatGPT 3.5, 4o, GPT-5-low, etc, I don't take anything they write seriously anymore.
Never, ever, use a cloud password manager, that's just dumb. Combining these things together in some sort of master account -- be it Google, Apple, Microsoft -- is also terrible. It's like leaving all of your savings accounts, checking, and investments at a single bank.
All of this stuff is going to get way worse because of AI. You'll be talking to real people you know personally who are 100% not AI but were tricked in to asking you to do something by other AI enabled scammers. However aggressive I've suggested people be in the past probably isn't going to be enough for 5 years from now.
These things have always been possible, and have been done, but now they can be done at scale, with advanced testing to figure out what works on who, whereas before it was targeting the guy who kept posting pictures of expensive watches on his public Instagram.
Whether its car thieves or drug dealers, these exist in the West today by explicit choice, not because it is impossible to stop.
Why are they idiots? Because western Europe is not yet authoritarian and thus there is little personal benefit to hasten a slide towards it, there are so many other ways to gain power in a free society. (I wouldn't bet money that Europe will remain free in 25 years.)
There is a secondary problem here -- anything that decreases the information security of European countries hands more power to the US and China (and to a lesser degree other nations with advanced infosec capabilities like Russia and Israel.) If you are European (I'm not) the first thing that should be done is investigate the people pushing this stuff.
I've been more neutral about this in the past, but the current and future integration of LLMs (and other ML models) into the base operating system will mean these mobile phones do less of what the user wants and more of what the user does not want.
Secondly, Apple is in the process of becoming an adtech company and will not provide an alternative to Google.
Thirdly, Google may be forced to divest Android and the mobile business. If so, the buyer is likely to be as bad, or worse, because they'll have to figure out how to pay for the whole thing.
The amount of legalism that's been brought in by both sides, Apple/Google and the regulators, layered in lies (we need to approve the software, register the developers, to protect the user from software), is divorced from the reality of the hardware-software relationship. This has led us down a path where everyone is debating the topics that Google, Apple, and revolving door regulators choose rather than the underlying reality.
There is a simple solution to all of this: Google and Apple should no longer be allowed to operate any sort of "App Store" or software distribution channel.
If you understand there are multiple models from multiple providers, some of those models are better at certain things than others, and how you can get those models to complete your tasks, you are in the top 1% (probably less) of LLM users.
The western internet as we knew it is dead. Privacy is dead, we already live in a post-Snowden panopticon. With multiple always-on microphones in every public room and often in private rooms too. HD cameras are everywhere. If you live in any major city, hundreds to thousands of hours of footage, which might contain you, is being uploaded for public view and AI training daily.
There have been other open source deathblow laws passed in the EU like the Cyber Resilience Act and the Product Liability Directive which have been repeatedly dismissed by other commenters on hn. Earlier stuff like the GDPR was dismissed too as only affecting big companies. The arguments in support of these laws have basically been you are small so you don't need to follow them. That seems like lot of disrespect for the EU's legal system but maybe it's well deserved.
It's only a matter of time until ID verification will be mandated even to make a post like this one on sites like hn. Western companies assisted authoritarian prison states in monitoring, censoring and controlling their citizens, when they should have been doing the exact opposite. Now it's really hard to argue that it isn't possible here.