Hallucinating Parrots on Tumblr

A collection of things that are bad (or occasionally good) about AI and large language models

[hallucinatingparrots]

How do AI detectors work? (And why shouldn’t you trust them?)

Ever wondered how AI detectors work — and why they don’t really work that well? Here’s an overview of them and some of the big issues that c

[lamajaoscura]

Is there an AI bubble — and is it about to pop?

The bill may be coming due for Silicon Valley’s huge investment in AI.

[hallucinatingparrots]

Fingers crossed

[hallucinatingparrots]

Terrible AI-deas: Aspect, the AI social media space that monetizes your solipsism

A social media system where everyone but you is a bot. It’s another Terrible AI-dea.

I found out about a social media app that's entirely populated by bots, intentionally.

I'm excited to share with you the first post in our series on Terrible AI-deas, looking at "Aspect", the social media hellscape that uses LLMs to create AI friends just for you

[hallucinatingparrots]

A prescient slide from an IBM presentation back in 1979. [source]

We have a bad habit of thinking that computers are infallible at everything just because they are (approximately) infallible at math. The world is not math, though. The most beautiful equation can still hide biases, errors, and uninterrogated assumptions.

Outsourcing decisions of almost any kind to a computer is a cowardly dodge, even more so with our modern black-box AIs that can't tell us why a decision was made.

[hallucinatingparrots]

I was reading through an AI-generated article on how to evade AI detection, and I was kind of fascinated by the explanation of why you'd want to evade detection.

The "maintain academic integrity" idea is a fascinating one, as it suggests that if you cheat but don't get caught, there's nothing wrong with cheating. What a strange worldview some people (or, maybe just some robots) hold.

[hallucinatingparrots]

“Wrong but plausible” is the worst kind of wrong for AI

Generally, it’s better to have a small error than a big error. But the upside to a big error is that it’s easier to notice than a small one.

Our first post on the Hallucinating Parrots blog. What kinds of errors should AIs generate?