Austen Tucker
banner
thearcades.bsky.social
Austen Tucker
@thearcades.bsky.social
Blind novelist, musician, dev, and general weirdo. I never turn down an excuse to play pinball. Chicago, IL TheArcades.me
It's also a TF rabbit hole waiting to happen, and IMO one of the most well-written disney movies of all time.

I fucking love this movie.
August 30, 2025 at 4:35 AM
I believe you're talking about the part where I disagreed with yoru assertion that OpenAI was the equivelant of leaving power tools out on the side of the road for free.

And I disagreed with the assertion that those were equivalent.

I mean, at this point, ship sailed. ¯\_(ツ)_/¯
August 28, 2025 at 3:03 PM
No, that's a lot of context to pile onto what I said, which was "this is a power tool, there are ways to use this responsibly." The no true scots play was yours - I simply offered my own experience.
August 28, 2025 at 2:54 PM
Would love to see examples of how OpenAI is advertising their use cases as human replacement. They’ve spoken of augmentation quite often, sure, but…
August 28, 2025 at 1:57 PM
Yeah, and I ran a circular saw when I was 7 while I helped my dad build a house. And while I was using it I didn’t try to intentionally bypass the safety devices meant to keep me from hurting myself so that I could then do so.
August 28, 2025 at 1:57 PM
I disagree but understand the logic! (I'm also that weirdo who wishes there were more Adventure Playgrounds where we handed tools to kids, but I digress :) )
August 28, 2025 at 1:33 PM
I still maintain that these are _tools_, not some sort of human analogue (yet). I believe the correct level of scrutiny is similar to that of power tools. Yes, they can hurt, maim, or kill, but they’re so useful that we learn to use them safely.
August 28, 2025 at 1:05 PM
Well, in response we now have widespread, pervasive, human-in-the-loop surveillance by OpenAI. We also have parental controls coming (cool!), but as someone who uses this software daily I’m unhappy

Maybe we should be more litigious, I don’t know. But I don’t like the secondary results here
August 28, 2025 at 1:05 PM
I've gone mad with power
August 23, 2025 at 10:55 PM
metr.org/blog/2025-03...

I’d encourage you to take a look at this. The problem is on the radar of AI scientists.
Measuring AI Ability to Complete Long Tasks
We propose measuring AI performance in terms of the *length* of tasks AI agents can complete. We show that this metric has been consistently exponentially increasing over the past 6 years, with a doub...
metr.org
August 19, 2025 at 1:14 PM