The knowledge it sucks makes it safer. You're under no illusions to trust it.

I'm going to do the poor man's short and just sell up before this all happens. Would be nice to be wrong but I'm not willing to risk what I suspect.

That sounds like both a problem for you and LLM providers. Perhaps a problem for those around you too...

Shotgun to the hard drive is probably effective.

I refuse to believe that. Chat tells me it only says those things to me!

Honestly I believe if LLMs were capable of self improving they'd already be doing it. Besides 8x code output is a horrific metric for success.

Can I have your autograph?

Is an absolute clusterfuck of a tool. Even simple things it does poorly at in my experience. Perhaps if they spent more time improving it rather than trying to bury reviews pointing out its flaws it'd be good.

It worries me greatly if it fails so many businesses that have tied themselves to it are going to fail too. Most of those secondary businesses will be the resellers flogging an interface that others will have incorporated as their own. It's going to be such a shit show. I'm somewhat tempted to invest everything I have into a companies like palantir as I suspect they'll be doing very well contracted to help quell the civil unrest caused by the formers fall out. It's a weird I'd love to be wrong about the bubble but I also kind of hope I'm right and generative AI takes a much smaller role going forward. It's too destructive to both the environment and the minds of so many people especially when you account for what it actually achieves.

Intriguing much lower that I'd have imagined. Thank you!

To pay us. You're welcome.

Now we'll see if its a bubble or not.

So it's working on a group vote to detect false information? Sounds expensive.

Yeah the autodesk fusion one is horrific for that. Spends a good 20 minutes responding with various versions of "executing script" or "searching for information" then finally times out with a "I don't know what you wanted could you reword it please". Genuinely frustrating. All I wanted was rounded corners haha!

If it can go "I don't know" rather than spew bollocks as fact it may actually be the turning point for AI that would get me to use it. At least for stuff I don't already know about already. Will have to try with things the previous one claimed to know despite being provably wrong.