Crypto collapse? Get in loser, we’re pivoting to AI
davidgerard.co.uk
davidgerard.co.uk
This however I thought was the substance of it:
> The VCs’ actual use case for AI is treating workers badly.
> The Writer’s Guild of America, a labor union representing writers for TV and film in the US, is on strike for better pay and conditions. One of the reasons is that studio executives are using the threat of AI against them. Writers think the plan is to get a chatbot to generate a low-quality script, which the writers are then paid less in worse conditions to fix. [Guardian]
> Executives at the National Eating Disorders Association replaced hotline workers with a chatbot four days after the workers unionized. “This is about union busting, plain and simple,” said one helpline associate. The bot then gave wrong and damaging advice to users of the service: “Every single thing Tessa suggested were things that led to the development of my eating disorder.” The service has backtracked on using the chatbot. [Vice; Labor Notes; Vice; Daily Dot]
> Digital blackface: instead of actually hiring black models, Levi’s thought it would be a great idea to take white models and alter the images to look like black people. Levi’s claimed it would increase diversity if they faked the diversity. One agency tried using AI to synthesize a suitably stereotypical “Black voice” instead of hiring an actual black voice actor. [Business Insider, archive]
Wow look at this thing that has the potential to do labor that previously required a human's time and effort. When you capture the value it means you can either do more or have more leisure. When you don't capture the value it means the treadmill of "being valuable to my employer" just got faster.
Why should VCs continue to "capture the value" of our leisure time and drive it towards zero? I understand that makes them lots of money, but why is that a good way to run society?
Consider the same reasoning in a different scenario:
I'm so tired of people complaining about their cancer. Yeah, your cells divide. When they divide the way they're intended to, you live a normal life. When they divide out of control, your health suffers. Thems the breaks.
I think we can all agree this argument is ridiculous. It describes the mechanism at work, acknowledges the harm that's taking place, but ultimately the argument here is "if we have an explanation for what's happening, then it isn't a problem, it's a fact of life." Clearly that's not so - just because we understand cancer, doesn't mean we accept it.
Help me understand why your argument is different?
Yes, but also it still does "require a human's time and effort", what else is the required training data for these models except a large amount of "human time and effort". It's just scraped for free and all munged together. Bulk, unpaid labour.
Even the ones left in crypto are basically trying to associate their crypto projects with AI as much as possible.
If real estate was hot, crypto would try to associate with real estate. If room temperature superconductors are invented, they'd say crypto for superconductors. Basically, in order for useless crypto shitcoins to get attention, they attach themselves to anything that's hot.
Alternatively, what you have said is true of all tools: a hammer does not need to be conscious to be dangerous, it just needs to be put in the hands of bad actors; well yeah, that is true of everything all the time...
And now just look at Sequoia. The bottom of the barrel with the worst takes in technology. How much a company can fall.
"Don’t want to worry anyone, but I just asked ChatGPT to build me a better paperclip."
Given the means, would it kill anyone who gets in the way of more paperclips?
This article seems more like a critique of AI doomerism.
Funny this peace could have been written by chatgpt with the same or better quality.
Only good thing is that crypto VC is now starting to invest in something worthy
Stopped reading after this part. To be so intentionally misleading and dismissive of truly miraculous work by the LLM community and the OpenAI team is not for me.
But our main applicable experience is as finance journalists writing up crypto grifters. And it's not only the same grift, but literally a lot of the same grifters.
It's never about the technology, it's always about the money.
I'm tired of these arguments. Very very few people are anthropomorphizing chatgpt.. very few. The majority of people both technical and non technical who have played with chatGPT in a non trivial way are aware of the chatbots limitations. It's like recognizing a crazy person on the street. Humans are well equipped for that.
This argument characterizes the average person as some kind of stupid buffoon as if he/she can't tell chatGPT really screws shit up. Sure there are a few gullible outliers but as a generality his claim is simply completely false. Pretty much everyone and I mean everyone is aware about the limitations of LLMs.
This is a weak and repeated trope that's being regurgitated as if the critics are LLMs themselves.
Let me specify exactly what's going on. People who are afraid of/support AI are more speaking to the potential of AI. Why? Because just as much as this thing hallucinates about half the time it answers a complex question with an equally complex and correct answer and that answer in isolation is often indistinguishable or even at times superior to a humans answer.
Yes we know it hallucinates and forgets shit. This is obvious, no need to readdress an obvious weakness that everyone is aware about. If critics want to have a real discussion then they seriously need to address the actual strengths and phenomenons of LLMs instead of repeatedly highlighting the obvious weaknesses.
Because while we have somewhat of an explanation for the hallucinations we currently don't know how chatGPT was able to do something like this:
https://www.engraved.blog/building-a-virtual-machine-inside/
Read to the end of you haven't seen this. The ending is what is quite unexplainable by experts. You can't just trivialize that entire post as if it was just a statistical phenomenon. There's obviously an alternative angle here.
The technology behind LLMs probably does have some interesting and valuable applications. But look at the hype and follow the money with a skeptical eye. If so-called AI does or will soon deliver real benefits then why all of the hype around it? Show, don't tell, and especially don't tell stories based on possible future advances or breakthroughs.
Future editions of Extraordinary Popular Delusions and the Madness of Crowds will have long chapters about crypto and AI. If you think most people understand the tech and its limitations and don't act according to gullibility, ignorance, and greed that book should disabuse you.
He believes we basically anthropomorphize each other. He made a great point how every child basically anthropomorphizes a doll. Anthropomorphizing is just second nature to us.
I would highly recommend reading his two longish medium essays on the subject.
The part about AI and Extraordinary Popular Delusions and the Madness of Crowds is just utter bullshit. Come on. I am sure you don't really believe that.
Look up the definition of forget. It is not a human exclusive action. Therefore it is not "anthropomorphizing": https://www.merriam-webster.com/dictionary/forget