FriendOfDeSoto

decent support chatbot

is an oxymoron.

Before we gloat too much - and let's be honest, we all wanna - CEOs tend to be of a certain vintage. I remember how I could program the VCR and my parents decidedly could not. Old folks' opinions may be less relevant here.

Bringing all these tools in is basically giving magic beans to cave people. How would they know how to use them effectively? All the while trying to figure out if they are indeed magic. This, sadly, could just be the anomaly before the numbers go up. This isn't proof positive that it's all horse shit just yet. It's just confirmation that the peddlers are overflowing with it.

Absolutely. Thiel-backed is a red flag hoisted on a flagpole whose tip reaches just below the ISS, whose width is at least 1000 football fields long, which is illuminated at all times by 5000 spotlights, and which is simultaneously and counterintuitively on fire despite the lack of oxygen at that altitude.

I'm also getting my tax advice from Gonzo on LinkedIn.

What a sad state of journalism. Celebrity posted something on social. Opinions are divided. Commenters said this. Others said that. Thanks for looking at our ads.

She may be right, I dunno. I'm criticizing the publication and quality of the story, not the actor and alleged AI influencer.

All the so-called AI companies are [expletive redacted] that often obtained their training data in questionable ways. And they should be sued and made to pay.

What is missing in my opinion from a lot of artists, who now hold one fist in the air and a pitchfork in the other marching on Silicon Valley, is an acknowledgement of risk that they took when they put their artwork themselves on the internet. The risk used to be other people could take, copy, duplicate it but this was balanced by being able to monetize it. That nothing that ends up online is ever safe was known two decades ago. If your stuff was stolen from galleries and coffee table books, I'm not talking about you. If you made your stuff available so search engines and social media sites could bring you costumers, you're it.

Also, why are companies able to obtain training data in "creative" ways? Because you and me and all of us like cutting corners and getting shit for free. So we steal, copy, duplicate, torrent stuff. Sure, point your finger at Facebook for torrenting together their model. Also point your finger at the people who provided the training data this way. Many people will end up pointing their fingers at themselves here.

I think the author invokes the Luddites and doesn't realize that we might just see history repeating. The Luddites were skilled artisans fighting against slave labor automization. They lost. Many were displaced by industrial progress, a smaller number remained. Fast forward to today and visual artists like the author fight the evil automated LLMs. They will also lose; a small number will prevail.

Every technical advancement has brought these upheavals and we are in one right now. There are far fewer landscape and portrait painters around today because they had to go Picasso or impressionist when photography rolled around. There are far fewer negative developer jobs these days as entry level jobs in photography because use of film has fallen off a cliff. We also have fewer manual typesetters and no cavalry to speak of. Shit changes. Shit is changing now.

Art will prevail. Human made one will be sought after. But there market will not be the same. The only thing we can do now is trying to catch all those people and jobs this technological leap will shuffle loose. Like we tried our best with coal miners or factory jobs that went to China.

I don't think the author likes mastodon dot social...

I couldn't get through all of this blog post because it's repeating the same point 500 times. I get the theoretical threat scenario they are painting; what's missing are the receipts. Is moderation on that instance actually getting worse? Have we talked with admins on the record how they don't dare defederate from that alleged wretched hive of scum and villainy? And two other angles are missing: (1) a name brand instance might be a good starting point on the fediverse. It's still better than Xwitter. And (2) people are not donating enough to their instances, who are then run on dedication and held together by duct tape. The fear of having one's instance shut down because the admin is out of money and duct tape makes people gravitate towards the bigger instances.

I'm not opposed to recommending people to find other, smaller instances - that is a good idea. It's just this blog post reads more like a hit piece.

I don't hate so-called AI per se. I see great use cases for people with disabilities. There are promising signs of it improving medical diagnoses (under properly tested conditions). I think even in my life I will learn to use some of the tools. Eventually. I try to avoid it right now as much as I can.

I hate the people peddling so-called AI as the solution to all problems, including already solved problems. I hate the mad rush on it because it risks negating all the positive greenhouse emission savings we have managed to get done. It will probably incur a greater water debt, i.e. more drinking water future generations will be forced to desalinate if they want to live. And it will make the next computing device you want to buy mad expensive because of the RAM shortage. I hate that this rush is a bubble that may not burst but drives prices up.

If I understand this correctly, what they need is an AI that can also drink a lot of beer! Solving two problems in one fell swoop. But where is Silicon Valley on this front? Nowhere. Sad!

This is sign that Google is worried that a market of 500 million people could decide to move away from the US tech giants. Very worried, judging by this flimsy fear-driven argument. Good.

I sympathize with your point of view here. I feel like that ship has sailed though. Messaging is the preferred means. That ship is not coming back any more.

Email is not well protected unless you and everybody communicating with you is taking extras precautions. Signal is E2E encrypted, WhatsApp also but owned by Meta so barf, Telegram's encryption status is complicated but probably better than plain email. There is a privacy advantage.

I treat instant messages that have the content of an email as such. I'll reply in my own time. Just because I got it instantly doesn't mean I need to act on it right away. I have some groups and contacts muted and have set quiet hours on my phone for evenings and nights. My advice is to look for ways to manage the stress you feel about this. That could mean going off the chat apps all together but I think you can also tweak settings and your behavior.

Two factors play into this. 1) This hype has been around long enough to have planned, built, and put online new data centers. So the bottlenecks some providers had in the early days have widened already. And 2) they need capacity for training more than using the models.