An important characteristic of a model is “stability.” Stability means that small changes in input produce small changes in output.

Stability is important for predictability. For instance, suppose you want to make a customer support portal. You add a bot hoping that it will guide the user to the desired workflow. You test the bot by asking it a bunch of variations of questions, probably with some RLHF. But then when it goes to production, people will start asking it variations of questions that you didn’t test (guaranteed). What you want ideally, is that it will map the variants to the best workflow that matches what the customer wants. Second best would be to say “I don’t know.” But what we have are bots who will just generate some crazy off-the-wall crap, and no way to prevent it.

Mozilla is going to kill it too. https://youtube.com/clip/UgkxsdmHepmqMNqWM0BmW2RHemnarYeQk0bk?feature=shared

I’ve been voting for a while and never have I seen a candidate on the ballot who was against capitalism.

Both CEOs are horrible but the new one is a former McKinsey consultant with a background in finance and the silicon-valley C-suite. According to statements she put out her strategy is: layoffs and AI.

Reality is about to get all melty and people are gonna have six fingers.

Is it really possible to sign away a right to sue a company, especially hidden in an EULA?

Yes it is. It is called “forced arbitration” and pretty much every contract you are compelled to sign has it.

In any kind of just society with a fair legal system it would not be legal. But that doesn’t describe us or our legal system.

My guess is you’re seeing the computer go into a reject loop until a human operator finally takes over.

My experience is that Firefox often has problems on Google-owned properties. Either performance/responsiveness or functionality just not working. Why this would be is left as an exercise for the reader.

Municipal police mostly came from the great railroad strike of 1871.

Funny you should mention that McKinsey published a paper a few months back concluding that GenAI will take over most of the jobs in America because it was good at doing what McKinsey Associates do. Missed by the authors is that the job of a McKinsey associate is to confidently spout nonsense all day long and that’s actually exactly what chatgpt is programmed to do.

American Psycho (Sam Altman) and his chorus have been hyping AI and the rest of the world’s reaction has ranged from “these guys seem smart and chatgpt is impressive so what do I know?” to “isn’t this guy a bitcoin bro?”

I feel like customer support is one place where AI may actually be used going forward because companies don’t really care if their customers get support. The only wrinkle is that if companies get held to promises the AI makes (there’s that Canada Air incident from last year where the AI offered a refund and the company tried to walk it back).

Naw if they’re publicly bashing it they’ve already dumped on all the downside risk onto their customers and now they’re net short.

Game of Life has cool emergent properties that are a lot more interesting and fun to play with than LLMs. LLMs also have emergent properties like, for instance, failing classification due to the manipulation of individual image pixels.

I suspect Intuit fired those workers for other reasons (free file) and are using AI as an excuse because to admit that free-file is an existential threat to their business is to admit that their company has no long term business prospects.

This is the same middlebrow dismissal that AI advocates have been using for years.

“It’s just a stochastic parrot.” “How do you know that you aren’t just a stochastic parrot?”

Well we do know. There are experts on human cognition. They have been studying it for decades. We may not know enough about it to know how to make a computer do it. But we certainly know enough about it to know when a computer chatbot is not doing it.

The space station’s orbit has been adjusted continuously over its lifetime initially by attaching a shuttle to it and doing a burn of the shuttle’s engines and later doing the same with progress modules.

My bet is the original expectation of the designers was to deorbit by attaching centaurs (or whatever) to the existing docking ports and rotate the beast to the right attitude for a deorbit burn.

NASA has more recently said they want the reentry to be as steep as possible to minimize the size of the debris field, and is using that to justify the development of a new specialized deorbit vehicle. No doubt SpaceX will declare that Starship is the proper vehicle for this, and then will plow the $800 million into the Starship program. The money they got for Artemus is already long gone and Starship has failed to demonstrate key components of the Artemus plan. Dear Moon has been cancelled so NASA and Artemus are the only customers they have left. NASA knows that without a cash injection Artemus is at risk.

Honestly this seems like a way to back-door inject another $800 million into the failing starship program.

I mean we’re sitting here on the lemmyverse having a conversation…

But yeah creators should upload to peertube but they won’t get any meaningful viewership there. The only way to break the network affect stranglehold google/youtube has over video content on the internet is making sure that if you do produce that content it’s available via other channels.

Yeah the fediverse has lower engagement all around because the community is a lot smaller. This is especially true in “long tail” communities. However, the upside is that there are no bots, dark patterns, or manipulated feeds.

That being said, while I appreciate the chronological feed I do wish there was some way to “weigh” less active communities so that I can see their activity in my feed without them being drowned out by the busier communities. I’ve noticed that I’ve gone to communities that I’m definitely subscribed to, and seen that there were several posts that I missed because the posts were drowned out by content in busy communities like, for instance, technology@beehaw.org