It shows that some measure of supervision is needed. That firing all of the coders, and unleashing bots is probably a bad idea (at least for the near future)
I dont know why people seem to care so much about llm safety. They’re trained on the internet. If you want to look up questionable stuff, it’s likely just a google search away
Suppose we have an LLM in an agentic loop, acting on your behalf, perhaps building code, or writing e-mails. Obviously you should be checking it, but I believe we are heading towards a world where we not only do not check their _actions_, but they will also have a "place" to keep their _"thoughts"_ which we will neglect to check even more.
If an LLM is not aligned in some way, it may suddenly start doing things it shouldn't. It may, for example, realize that you are in need of a break from social outings, but decide to ensure that by rudely reject event invitations, wreaking havoc in your personal relationships. It may see that you are in need of money and resort to somehow scamming people.
Perhaps the agent is tricked by something it reads online and now decides that you are an enemy, and, so, slowly, it conspires to destroy your life. If it can control your house appliances, perhaps it does something to keep you inside or, worse, to actually hurt you.
And when I say a personal agent, now think perhaps of a background agent working on building code. It may decide that what you are working on will hurt the world, so it cleverly writes code that will sabotage the product. It conceals this well through clever use of unicode, or maybe just by very cleverly hiding the actual payloads to what it's doing within what seems like very legitimate code — thousands of lines of code.
This may seem like science fiction, but if you actually think about it for a while, it really isn't. It's a very real scenario that we're heading very fast towards.
I will concede that perhaps the problems I am describing transcend the issue of alignment, but I do think that research into alignment is essential to ensure we can work on these specific issues.
Note that this does not mean I am against uncensored models. I think uncensored/"unaligned" models are essential. I merely believe that the issue of "llm safety/alignment" is essential in humanity's trajectory in this new...."transhuman" or "post-human" path.
I dont know why people seem to care so much about llm safety.
That's kind of an odd question?
To me it's obvious that people want to make money. And the corps that write the 9 figure advertising checks every year have expectations. Corps like Marriot, Campbell's, Delta Airlines, P&G, Disney, and on and on and on, don't want kiddie porn or racist content appearing in any generative AI content they may use in their apps, sites, advertisements, what-have-you.
In simplistic terms, demonstrably safe LLM's equals mountains of money. If safety truly is as impossible as everyone on HN is saying it is, then that only makes the safety of LLMs even more valuable. Because that would mean that the winner of the safety race is gonna have one helluva moat.
It was initially drummed up as a play to create a regulation moat. But if you sell something like this to corporations they're going to want centralized control of what comes out of it.
It’s interesting. I think the real way to do this is gradually scale up. Crossing the ocean is hard mode. Instead start by something much shorter and land based. Then you at least have a stable platform to work on and can focus on the other hard problems
In 1st world countries, land based cables often cost more because you have thousands of people along the route who all don't want a power line through their small village.
The thing is, its not just Meta. It's anyone with a square inch of pixels that you can slap an ad on. Google, Amazon, Microsoft, etc. Meta gets a lot of flack, but it's easy to stop using Facebook. It's a lot harder to stop using Google or Amazon.
> it's easy to stop using Facebook. It's a lot harder to stop using Google or Amazon.
Not in my country, where Facebook owns basically the entire messaging space through Messenger, Instagram, and WhatsApp. I have been able to switch my default search engine, email provider, mobile OS, office suite, etc. from Google no problem, but the network effect has been too strong to get most of my friends out of Meta's ad network. It doesn't help that most groups and basically all events are organized through there.
The only real messaging alternative with significant adoption at this point seems to be Telegram, which I cannot support for numerous reasons, primarily for the founder selling the entire user base of vkontakte to an authoritarian state.
One thing I have noticed is that local businesses don't have websites any more, they have Facebook pages instead. So if you want to find out about local events, Facebook is the only way.
As sibling said, using the mautrix bridges. I started with just bridge mode I think, and then switched to puppet mode, which seemed to integrate much better. Pictures and everything work fine.
However, you have to maintain a homeserver, which is not for everyone. If wouldn't have one already, I would just pay the element.io service which includes bridging (but I haven't tried it out).
Over the course of 5-6 years, I managed to move everyone that is important enough to me to Telegram. I don't care about privacy at all, I just couldn't stand the wrong abrasive spit-in-your-face whatsapp UX. It's funny how facebook was enshitified long before it was cool.
Now I don't have notifications enabled for Whatsapp and I rarely answer messages there. It makes me a freak by many standards but I'm free from enshitification in this regard!
They came here perhaps two years ago and everybody I know agrees they suck and can not compete with our local alternatives. They're not cheaper and the website is an utter mess of auto translations and dark patterns. Order flash storage and pray it's not counterfeit. We use Prisjakt[1] to compare prices and price history between shops.
About Google... Yes, YouTube is hard for me to avoid. I use StartPage for search (which is Google search proxy just like DDG is a Bing proxy). Switching to DDG wouldn't be that much of a sacrifice. I don't have a Google account and YT works well via RSS. FreshRSS is basically my YT frontend.
Facebook is the hardest to leave, without a doubt. Civil society isn't organized around any of Amazon's or Google's services. Parent group for school? Facebook. A group for your neighborhood? Facebook. Local trading of goods? Facebook marketplace. I've made the data takeout and have been hovering over that account deletion button I don't know how many times... But I don't follow through.
I'm reasonably sure that if you will block every single Amazon owned address at your entry firewall, a majority of the apps or sites will stop working for you. That's what is implied by saying that it is not possible to avoid Google, Amazon and either one of the MS or Apple.
Kagi has a great browser, proton has mail and calendars (I mean everything has calendars, there isn’t really much lock in there), maps are (location dependent) very fungible.
The OS and documents are trickier as I’ve never used either but I did only say 70%. The thing I would find hardest would be converting sign in with google accounts
If the average person was so inclined, they could probably degoogle themselves 70% in one day, 90% in a month, and 98% in a year with very little cognitive load
If anyone is interested, there was an experiment about that. A week without either of Google, Amazon, Microsoft, Apple, Facebook. And last week without all 5. Very illuminating.
This is really about cost-cutting. If Amazon felt it were about being spread too thin, the obvious solution is to make a short list of charities. Or increase the amount donated.
For all the love DDG gets on this site, it’s not really an independent search engine. It relies heavily on the Bing index, despite spinning it as using it as one of many signals. Brave has its own index and that puts it in another class. In my experience the results have been higher quality, too.
Have you spent much time with Brave Search? I use Firefox and currently have DDD as the default search on my laptop, but I often use the !g operator since the DDD results are routinely not as good as Google. I’ve had a significantly better experience with Brave Search, though. I’ll probably switch my default soon, once I learn the operators that Brave supports.
The bang operators for ddg justify it as my default even knowing that it isn’t very good for some queries. The habit of being able to reroute a search to the engine that will handle it most appropriately is a killer feature. Only drawback is that the redirect can take an extra few seconds sometimes.
It’s hard to imagine replacing it. If Brave search becomes excellent then I can just use the bang for it.
What wrong if they’re trying to find more money while providing privacy?
Would you rather have a browser from rich monopoly like google? Or an alternative thats self sustainable with its own money?
Of course it does. There is no such thing as a free lunch, and companies have to make money. One of the best things about Brave is that it's aggressively trying to build independent revenue streams and a footprint on the net. A company with solid, diverse, independent sources of revenue is much better positioned to keep making independent decisions.
Was it a PR move to do spans the PR department? “Tesla, where every dollar goes towards making your car better!”. Or did they seriously disband their PR department!?
Apple is trying to see what the reaction is. They've got some negative press recently for pricing things a little too high (say a $1k monitor stand or $700 wheels). They're watching to see what the reaction to the rumor is. If people say 'no one would ever pay that', then they'll dial it back a bit. If the reaction is 'eh, maybe for those specs', then full steam ahead.
Either way, they can just deny the leak as a rumor without any basis.
Seems like a conspiracy theory. Apple has years of experience creating products that people love at all different price points (original iPhone was crazy expensive, for example).
If the product is legitimately better than everything else on the market, it doesn't matter how expensive it is on initial launch.
The original iPhone was more expensive compared to other phones but it was not priced like an Apple hardware product for professionals. Between 500$ and 600$. 600$ is about 780$ in today dollars.
From the article, it would have "an outward display that could be used to show content to people near you or to check information when the headset is not on your head"
That detail seems a little suspect to me in this whole story
