Are you not using AI at all? I still think this is a bubble about to burst, but also, they each do have actual groundbreaking products. Doesn’t mean I believe Altman or OpenAI will necessarily be around in 15 years, but in my opinion they’re more of a WebVan than they are anything.
They are conning people by selling the lie that they are replacing the working class, that I agree with. But they still have something that will stick around, even if it ends up being less loud in our lives. The bubble will burst, but it won’t disappear.
Live low stakes low-power translation of common sentences on my phone is pretty fucking cool.
They’re going to be fantastic for organizing the next internal genocide in the US, they’re perfect for mass surveillance and targeting.
They are magnificent accountability sinks for people who want to blow up a school full of children, deny medical care, etc.
Meanwhile I’m only seeing the quality of software plummet at MSFT and Amazon and Google and Meta. I would have expected them to have fewer things fall over and catch fire if this did was so good
I think a lot of people are just in love with the chatbot that tells them they’re smart and love the idea of having a little homunculus that loves them to boss around.
It freaks me out to see people who are allegedly working on complex engineering projects look to their tools for emotional support.
I look at chatbots like I look at speed, I know people who graduated from college by staying awake for days on meth. Can’t deny that they have a diploma now, but it makes you crazy.
I don’t have a project that needs to use an llm and, frankly, I don’t want to end up like Steve Yegge or Satya Nadella or Kent Overstreet. They’ve all clearly got a screw loose now, and it’s pretty clear that chatbots had a lot to do with it.
If I am going to use an llm, I’m going to use it as a component that I interact with programmatically to find approximate solutions to a nasty space.
Using one to generate code listings that your roof depends on seems fucking insane to me. Using one to write developer documentation seems fucking insane to me.
It’s a cargo cult imo. Yes, the radio was real. But having a radio does not mean you have an air force on the other side. Right now the hyperscalers are confidently calling out orders for chocolate and mayonnaise and bullets and antibiotics, confident the planes are coming.
Yes you can create a working program with an llm. I think in the end, though, a recently-fired engineer with a mortgage will be cheaper and more available than Claude.
I agree that people that dismiss current AI models as hype and slop have not used them to any extent. I am constantly astounded by their abilities. The problem with dismissing them as a grift is that it is a distraction from how we should be thinking clearly about how to confine them in the event that they are not a grift.
They are, at best, an unreliable tool you need to babysit. And because of that, you need to be able to do what you’re asking the tool to do so you can verify it actually did what you asked.
That’s is not at all how these tools are being marketed, or how many are using them. So the current result is 90% of usage is hype and slop.
Nah, just build a harness that validates the output of one model by running it through the same model again to check for hallucinations… And to make sure that second pass isn’t hallucinating, uh… run it through a model a third time to check the second isn’t hallucinating.
Are you not using AI at all? I still think this is a bubble about to burst, but also, they each do have actual groundbreaking products. Doesn’t mean I believe Altman or OpenAI will necessarily be around in 15 years, but in my opinion they’re more of a WebVan than they are anything.
They are conning people by selling the lie that they are replacing the working class, that I agree with. But they still have something that will stick around, even if it ends up being less loud in our lives. The bubble will burst, but it won’t disappear.
Live low stakes low-power translation of common sentences on my phone is pretty fucking cool.
They’re going to be fantastic for organizing the next internal genocide in the US, they’re perfect for mass surveillance and targeting.
They are magnificent accountability sinks for people who want to blow up a school full of children, deny medical care, etc.
Meanwhile I’m only seeing the quality of software plummet at MSFT and Amazon and Google and Meta. I would have expected them to have fewer things fall over and catch fire if this did was so good
I think a lot of people are just in love with the chatbot that tells them they’re smart and love the idea of having a little homunculus that loves them to boss around.
It freaks me out to see people who are allegedly working on complex engineering projects look to their tools for emotional support.
I look at chatbots like I look at speed, I know people who graduated from college by staying awake for days on meth. Can’t deny that they have a diploma now, but it makes you crazy.
I don’t have a project that needs to use an llm and, frankly, I don’t want to end up like Steve Yegge or Satya Nadella or Kent Overstreet. They’ve all clearly got a screw loose now, and it’s pretty clear that chatbots had a lot to do with it.
If I am going to use an llm, I’m going to use it as a component that I interact with programmatically to find approximate solutions to a nasty space.
Using one to generate code listings that your roof depends on seems fucking insane to me. Using one to write developer documentation seems fucking insane to me.
It’s a cargo cult imo. Yes, the radio was real. But having a radio does not mean you have an air force on the other side. Right now the hyperscalers are confidently calling out orders for chocolate and mayonnaise and bullets and antibiotics, confident the planes are coming.
Yes you can create a working program with an llm. I think in the end, though, a recently-fired engineer with a mortgage will be cheaper and more available than Claude.
I agree that people that dismiss current AI models as hype and slop have not used them to any extent. I am constantly astounded by their abilities. The problem with dismissing them as a grift is that it is a distraction from how we should be thinking clearly about how to confine them in the event that they are not a grift.
They are, at best, an unreliable tool you need to babysit. And because of that, you need to be able to do what you’re asking the tool to do so you can verify it actually did what you asked.
That’s is not at all how these tools are being marketed, or how many are using them. So the current result is 90% of usage is hype and slop.
Nah, just build a harness that validates the output of one model by running it through the same model again to check for hallucinations… And to make sure that second pass isn’t hallucinating, uh… run it through a model a third time to check the second isn’t hallucinating.
/s
Surely it will get better each time! Just like photocopying a copy!
Only by constantly re-photocopying something can something truly timeless ever be reified