90s Dad

[90s Dad @inflatable_josh]

@LostKimber @8biteric

[90s Dad @inflatable_josh]

@neonlantern148 @JayHooft You do know you don't have to be a massive fucking dork, right?

[90s Dad @inflatable_josh]

@badazn its funny because he is in a lot of financial trouble now and its basically because he cant read or write

[90s Dad @inflatable_josh]

@BriaNamtar i think its the hypnotic dance moves

[90s Dad @inflatable_josh]

@M_bolaji__ @LavenderGhast indian myna

[90s Dad @inflatable_josh]

@MannyFEstez @8biteric epic game. they need to remake this

[90s Dad @inflatable_josh]

@8biteric

[90s Dad @inflatable_josh]

@DudespostingWs the Lyrebird is his final boss. it will be his undoing.

[90s Dad @inflatable_josh]

@beautyskillz Call of Duty - Dad Ops

[90s Dad @inflatable_josh]

This is the reasoning why - Because 17 has become a kind of “human-looking random number”, and LLMs inherit that bias from human-generated text.

When you say “pick a number between 1 and 30,” an LLM usually isn’t running a true random-number generator. It is predicting which token is a plausible continuation. Numbers like 1, 10, 15, 20, 25, 30 feel too obvious, while numbers near the edges feel less “random.” 17 has several properties that make it unusually attractive:

It’s near the middle, but not exactly the midpoint.

It’s odd.

It’s prime.

It doesn’t end in 0 or 5.

Humans themselves disproportionately choose numbers like 7, 17, 23, and 27 when asked to “pick randomly.”

Training data contains countless examples where people use 17 specifically as an arbitrary or supposedly random number.

So models learn a statistical association something like:

> “When someone asks for a random-looking integer, 17 is a very plausible answer.”

There’s also a reinforcing effect: people have noticed the phenomenon, posted about “ChatGPT always picking 17,” tested models with it, and generated even more text associating that prompt with 17. That can further strengthen the pattern in later training data.

It’s similar to the famous “pick a number from 1–10” → 7 effect in humans. Seven feels random because it avoids the obvious choices. For 1–30, 17 occupies a similar psychological sweet spot.

[90s Dad @inflatable_josh]

@GoblinWithAPlan @CreagenDow I'm not. I don't pay for that shit. I put it through grok and chatgpt and they said 17 but I thought Geminis response was funny

[90s Dad @inflatable_josh]

@gtainfobr He doesn't have that function in the build that he is using. He is using a small slice of the early game that was used in one of the builds for testing. The areas he is showing will likely be completely different come launch.

[90s Dad @inflatable_josh]

A build doesnt mean full game. It's likely he has a slice of the game that was intended for testing. There's probably a reason why his videos are all around the same length and why he hasn't got access to both characters. The build that he has is probably a slice of an early stage of the game. I've seen enough. I don't care to see any more spoilers. I'll wait until the game comes out to enjoy it

[Papa Heme @Papa_Heme]

Say what you will about Silverchair but this song is unreal. One of the best to come out of the 90s.

[90s Dad @inflatable_josh]

@skyecorner96 This is what Hollywood thinks dudes want their home to look like

[90s Dad @inflatable_josh]

@swiftism100 Nobody did that you fucking fuck!

[90s Dad @inflatable_josh]

@ElofsonJess I'm starting to think he is Death

[90s Dad @inflatable_josh]

@GTA6_Premium Private detectives and investigators usually don't tell the person they're investigating who their client is. It is faker than fake

[90s Dad @inflatable_josh]

@videotech a private detective wouldnt say who he is working for