This Blog is Anti AI Slop

devoted to pointing out ai so other people don't fall for it

[klfg]

[source]

[klfg]

🚨BREAKING: OpenAI published a paper proving that ChatGPT will always make things up.

Not sometimes. Not until the next update. Always. They proved it with math.

Even with perfect training data and unlimited computing power, AI models will still confidently tell you things that are completely false. This isn't a bug they're working on. It's baked into how these systems work at a fundamental level.

And their own numbers are brutal. OpenAI's o1 reasoning model hallucinates 16% of the time. Their newer o3 model? 33%. Their newest o4-mini? 48%. Nearly half of what their most recent model tells you could be fabricated. The "smarter" models are actually getting worse at telling the truth.

Here's why it can't be fixed. Language models work by predicting the next word based on probability. When they hit something uncertain, they don't pause. They don't flag it. They guess. And they guess with complete confidence, because that's exactly what they were trained to do.

The researchers looked at the 10 biggest AI benchmarks used to measure how good these models are. 9 out of 10 give the same score for saying "I don't know" as for giving a completely wrong answer: zero points. The entire testing system literally punishes honesty and rewards guessing.

So the AI learned the optimal strategy: always guess. Never admit uncertainty. Sound confident even when you're making it up.

OpenAI's proposed fix? Have ChatGPT say "I don't know" when it's unsure. Their own math shows this would mean roughly 30% of your questions get no answer. Imagine asking ChatGPT something three times out of ten and getting "I'm not confident enough to respond." Users would leave overnight. So the fix exists, but it would kill the product.

This isn't just OpenAI's problem. DeepMind and Tsinghua University independently reached the same conclusion. Three of the world's top AI labs, working separately, all agree: this is permanent.

Every time ChatGPT gives you an answer, ask yourself: is this real, or is it just a confident guess?

[embrace-monkey-brain]

Why language models hallucinate

OpenAI’s new research explains why language models hallucinate. The findings show how improved evaluations can enhance AI reliability, hones

Here's a good article explaining why it's all LLMs, not just chatGPT:

[calling-it-like-i-see-it]

I wanted to make a post. I've seen this post making the rounds but it is not tagged for AI. It is AI. The poster includes a reference photo to make you think it is made by a person. It is not.

IT IS AI SLOP

But how can you tell its AI? From far away it looks good. Its adorable its cute. I myself was almost tempted to reblog is just scrolling through. I wanted to look at the artists other works and there is none. This should be a dead giveaway because you aren't amazing at art and only post one thing.

If you aren't checking out who you reblog from (and trust me its hard to do so especially when you want to just scroll and enjoy your time, not play detective) I recommend just stopping and looking at something if it seems too good to be true. Looking closer at details you can see stuff that becomes impossible to ignore.

His eyes are closed. But you can only see one, the other is whited out/ not there.

His cape randomly has one tassle. While adding details can be a style choice sometimes, it is out of the blue and most people do not do this. Its weird. Its illogical. But maybe someone who does not understand what something should actually be, but just knows make it look like these other pictures that are old timey and have tassles would put it there (a certain program? You see where I am going?)

Other things leap out the more you question it.

The box is the largest offense. It is not a box. AI cannot understand why its failing. The inside has stripes and ledges. Look at the lid its morphing into the box. AI cannot tell where one object begins and the other ends. It doesnt even recess inwards it just creates another plane.

His boots aren't laced properly. They dont even have eyelets for the lace to go through.

His hand is morphing into the ribbon he's holding. His glove doesn't know where to end and created the lines on his thumb.

If you dont believe anything else I point out, please look at the Damian's name tag in the same picture. There are 4 dots. 2 for a colon, one for the i, and one just floating. AI does not know how to write. It does not know because it cannot think nor learn. It can only copy shapes.

Please for the love of God tag if you use AI. I do not endorse using AI (for mental, emotional, and environmental reasons) but I am not going to soapbox you if you do. I am going to call it out if you don't TAG IT! I am ridiculously frustrated with it and have the tag blocked for a reason. I want work by humans for humans. And if you don't respect others enough to tag for it, then I do not respect you enough not to call out bs.