Thread 58138711

Why aren't Pokemon Gold & Silver ever used in experiments and benchmarks? Aren't they supposed to be an upgrade?

>>58138711 (OP)
Because that begs the question of why not the next gen based on your same reasoning all the way to SV

Would be nice if the Claude Plays Pokemon channel swapped to Gold. Rather tired of Kanto and as long as the fucking chatbot has access to its navigator it will never beat the Rocket Hideout.

>>58138715
At a certain point, surely the complexity becomes too overwhelming for this stage?

>>58138718
>Rather tired of Kanto
>it should move on to Kanto
?

>>58138721
That's why the 1st gen is always a good benchmark, since where you stop as you go forward kind of gives away the answer because of the suspicion on its specificity, which kills some of the intrigue in caring about stuff like this

Has AI beaten pokemon yet or are they literally stupider than Twitch chat? Took them like 2 weeks despite being nothing but chaos

>>58138739
Gemini beat RB in 813 hours but with supplementary AI tools it had access to
Twitch Plays Pokemon beat it faster

>>58138711 (OP)
>Aren't they supposed to be an upgrade?
That's the problem.
Pokemon Red & Blue are extremely simple, yet appear complex. It's the perfect benchmark to impress investors.

I would bet 5 entire dollars that GPT-5 wouldn't even be able to enter Kanto; nevermind get to Mt. Silver.

>>58138739
>>58138745
Hilarious the monkeys at typewriter metaphor actually still provides more universal progress than having the machine attempt to do it by itself. It's like a giga autist because it becomes more obsessed with RNG manipulation and attempting to succeed at irrelevant things than progressing the game itself

>>58138734
Is your retardation learned or were you born stupid?

>>58138750
>the monkeys at typewriter metaphor actually still provides more universal progress
Wasn't there an "RNG plays Pokemon" stream that tried to play Ruby and it couldn't even leave Littleroot Town?

>>58138751
?
why are you posting your diary entries here?

>>58138764
To be fair, you have to have a very high IQ to understand Ruby and Sapphire.

>>58138770
So not only are you stupid, you're schizophrenic too. What a winning combination.

>>58138770
Masuda was never good at region design. That's the actual reason why all Pokemon-related projects start with Kanto

>>58138793
I love getting Ralts and Lotad (and Taillow) in the first half hour next to Oldale. I love beginning hunting for buried items starting in Petalburg. I love wondering if I should pick up Slakoth in the forest. I love getting the watering pail and Bullet Seed, having my first Doubles battles, and walking across the reflective pond on the bridge. I love riding in Mr. Briney's boat to Dewford and Slateport and walking along their beaches, and going back to explore the Abandoned Ship and the Regi cave.

I could go on, but the point is RSE is fun pretty much from start to finish. It's a constant adventure with lots of rewarding beats.

>>58138814
>I love linear game design because it gives me good mons off the bat to make me "feel like I'm exploring"

And you wonder why the games were all straight lines after Sinnoh

>>58138818
Hoenn has the best exploration of the 4 gens though.

>>58138869
As I posted >>58138793
Half the game is fun, the other half isn't. But it's still linear. There's no true freedom like Kanto, it's the same issues Johto has but they don't pretend that there's "options" which fucks the level curve. You can pick up a few junk items or shitmons like Jigglypuff

You know what's fun exploration? Beating Koga after Misty and getting Surf immediately to have access to the rest of the game. That shit is more fun than any of the later Pokemon games

>>58138711 (OP)
let them beat gen 1 first, it just makes sense to have ai beat them in order. claude is the only one that matters because it's actually trying to play without being given a minimap and exactly where to go. I'm sure they'll do gen 2 after it finally beats gen 1.