671–680 of 872 posts
Earlier quoted context omitted.
I pay for the subscription and don’t see any ads. It comes with YouTube Music. It’s great.
None of your videos have in-video ads? "This segment is sponsored by NordVPN!" style stuff?
Earlier quoted context omitted.
> Yep. Dark patterns you can see are not that dark by comparison, we will need another word for coming dark patterns disguised in llm responses As someone else said, you can probably filter responses through a small purpose-built/trained LLM that strips away dark patterns. If you start getting mostly empty responses as a result, then there was no value prior to the stripping anyway.
If you can't tell when a big expensive llm is subliminally grooming you to like/dislike something or is selective with information then another (probably small and cheaper) llm somehow can? Arms race?
this is already there and in prod but called AI "safety" (really corporate brand safety). The largest LLMs have already been shown to favor certain political parties based on the preferences of the group doing the training as well. Even technical people who should know better naively trust the response of an LLM well enough to allow to make API calls on their behalf. What would prevent an LLM provider to train their model to learn and manipulate an API to favor them or a "trusted partner" in some way? It's just like in the early days, "it's on the Internet, it has to be true".
Earlier quoted context omitted.
Don't ship fundamentally broken products would be step one for me. Sadly, there's a lot of people who are really excited about things that only occasionally work.
Lots of things only occasionally work but are still very useful. Google search for example. Would you say "pah why are you shipping a search engine that only sometimes finds what I'm looking for?"?
If there's nothing answering what I was looking for, I might try again with synonyms, or the think documents aren't indexed, or they don't exist.
That's a very different failure mode than blatantly lying to me. By lying to me, I'm not blaming myself, I'm blaming the AI.
Here is the experience when clicking a link on mobile: * Page loads, immediately when I start scrolling and reading a popup trying to get tracking consent * If I am lucky, there is a "necessary only". When unlucky I need to click "manage options" and first see how to reject all tracking * There is a sticky banner on top/bottom taking 20-30% of my screen upselling me a subscription or asking me to install their app. U…
For those who want to experience it: https://how-i-experience-web-today.com/ The only inaccurate thing of that meme page is that you only need to uncheck 5 cookie "partners", when in reality there should be at least a few hundred.
Earlier quoted context omitted.
> am I standing at the gates of hell in a line that includes Tower Records and Blockbuster? Maybe, but there’s a big difference - Netflix doesn’t rely on Blockbuster, and Spotify doesn’t need Tower Records. Google AI results do need your articles, and it returns the content of them to your readers without sending you the traffic. And Google is just trying to fend off ChatGPT and Meta and others, who absolutely will,…
The recent taking of people's content for AI training might be the most blatant example of rich well connected people having different rules in our society that I've ever witnessed. If a random person copied mass amounts of IP and resold it in a different product with zero attribution or compensation, and that product directly undercut the business of those same IP producers, they would be thrown in jail. Normal peop…
I've written high-quality technical how-tos for many years, starting with PC World magazine articles (supported by ads), a book that helped people learn Ruby on Rails (sales via Amazon), and more recently a website that's good for queries like "uninstall Homebrew" or "xcode command line tools" (sponsored by a carefully chosen advertiser). With both a (small) financial incentive and the intrinsic satisfaction of doing…
> am I standing at the gates of hell in a line that includes Tower Records and Blockbuster? Maybe, but there’s a big difference - Netflix doesn’t rely on Blockbuster, and Spotify doesn’t need Tower Records. Google AI results do need your articles, and it returns the content of them to your readers without sending you the traffic. And Google is just trying to fend off ChatGPT and Meta and others, who absolutely will,…
Even when you have them dead to rights (like with the Whisper hallucinations) the legal argument is hard to make. Besides, the defendants have unfathomable resources.
Earlier quoted context omitted.
Dear lord please don’t use an AI overview answer for food safety. If you made a bet with your friend and are using the AI overview to settle it, fine. But please please click on an actual result from a trusted source if you’re deciding what temperature to cook meat to
Mmm, I see this cutting both ways -- generally, I'd agree; safety critical things should not be left to an AI. However, cooking temperatures are information that has a factual ground truth (or at least one that has been decided on), has VERY broad distribution on the internet, and generally is a single, short "kernel" of information that has become subject to slop-ifying and "here's an article when you're looking for…
I don’t think this is substantively different from cooking temperature, so I’m not trusting that either.
Earlier quoted context omitted.
You forgot the part about when you actually get to the content, there's usually about 5 paragraphs of SEO filler text before it actually gets onto answering the topic of the post.
Big Mama's Best Brownie Recipe. Let's start at the beginning. I was born in 1956 in Chicago. My mother was a cruel drunk and the only thing my father hated more than his work was his family.
Earlier quoted context omitted.
The race to continue training and researching, however, is drive by competition that will fall away if competitors also can't raise more money to subsidise it. At that point the market may consolidate and progress slow, but not all providers will disappear - there are enough good models that can be hosted and served profitably indefinitely.
Seems like there can never be good enough models; the user will want it up-to-date models with respect to news and culture.
Even for the uses where it does matter, unless providers get squeezed down to zero margin, it's not that new models will never happen, but that the speed at which they can afford to produce large new models will slow.
Earlier quoted context omitted.
The troll gives me the main idea without having to find five tiny x's on the screen like some sadistic minigame then paywall me. I'll take the troll.
They dont. Most commenters react to the title and preexistent opinions. Rhey frequently misinterpret the article too - misconstructing arguments they dont like and such.