froggoboom

founder

[froggoboom @froggoboom]

Whats the point of #AI if our thoughts just get drowned in the noise?

[el.cine @EHuanglu]

It's over.

Google Veo 2 just wiped out the competition and became the new king of the AI media world.

No one can tell this is AI now.

10 wild examples:

[froggoboom @froggoboom]

@GjMcGowan essentially an interactive coupon code- there are a number of ways the LLM can be made to always refuse lowball offers, such as validation at checkout. I would be surprised if the final price output given by the LLM is plugged directly into their payment systems

[Omar Sanseviero @osanseviero]

Falcon 180B is out🤯

- 180B params
- Trained on 3.5 trillion tokens+7 million GPU hours
- Quality on par with PaLM 2 outperforms Llama 2 and GPT-3.5 across 13 benchmarks
- 4bit and 8bit precision with similar quality

Demo: https://t.co/JCMYid7VYn
Blog: https://t.co/vb8ppRHgJ5

[Garry Tan @garrytan]

Sad to hear Steam is decel

“No AI in games” is idiotic

[Andrej Karpathy @karpathy]

This is a baby GPT with two tokens 0/1 and context length of 3, viewing it as a finite state markov chain. It was trained on the sequence "111101111011110" for 50 iterations. The parameters and the architecture of the Transformer modifies the probabilities on the arrows.

E.g. we can see that:
- state 101 deterministically transitions to 011 in the training data, so the probability of that transition becomes higher (79%). Not near 100% because we only did 50 steps of optimization.
- state 111 goes to 111 and 110 with 50% probability each, which the model almost learns (45%, 55%).
- states like 000 are never encountered during training, but have relatively sharp transition probabilities, e.g. 73% of going to 001. This is a consequence of inductive biases in the Transformer. One might imagine wanting this to be 50%, except in a real deployment almost every input sequence is unique, not present in the training data verbatim.

Not really sure where I was going with this :D, I think it's interesting to train/study tiny GPTs because it becomes tractable to visualize and get an intuitive sense of the entire dynamical system. Play with here: https://t.co/8jdceMLpqy

[froggoboom @froggoboom]

@NotGhoulish froggoboom approves

[froggoboom @froggoboom]

@Hammybh https://t.co/ZqyojQwAXL

[Joscha Bach @Plinz]

in the future, we will not learn how to read and write, but how to deeply interpret seven plus/minus two dimensional embedding spaces, and by projecting the neural encoding back into the machine via neuralink, we will all learn how to be one with the machine

[froggoboom @froggoboom]

@NotGhoulish @greenocool1 ur the pillow

[froggoboom @froggoboom]

@yezeehw poetic tweet bro im wheezeing

[froggoboom @froggoboom]

@greenocool1 🤠

[froggoboom @froggoboom]

@justLukass @AlterationNFT @madewithunity amazing, i look forwards to seeing the hovercraft in the finished game! 🤠🤠

[IndianBrawlhallaCommunity @IBC_Brawlhalla]

⚔️𝗕𝗿𝗮𝘄𝗹𝗵𝗮𝗹𝗹𝗮 𝘅 𝗖𝗮𝘀𝘁𝗹𝗲𝘃𝗮𝗻𝗶𝗮 𝗖𝗿𝗼𝘀𝘀𝗼𝘃𝗲𝗿 𝗚𝗶𝘃𝗲𝗮𝘄𝗮𝘆! ⚔️

x1 Alucard (Ezio)
x1 Simon Belmont (Jhala)

✨2 Winners will be drawn in 5 days from now!

✨To Participate:

1. Follow @IBC_Brawlhalla
2. Like and Retweet
#Brawlhalla #Brawlhallagiveaway

[Ilya Abugov - Helika @AbugovIlya]

NPCs populate the #Metaverse so join #WaypointWeb3 to learn about how they are evolving in #web3 with:

@BraxtonTBragg @FuzzleOfficial
@519Odin @AlterationNFT
@creaoy @crossverse_xvrs
@dorukmoruq @PlayEldarune
@decimated_game

Don’t miss it: https://t.co/NEAEczgs0S

[froggoboom @froggoboom]

@creqr im thinking of switching mains from lin fei to ember 😅

[Goober @GooberMVS]

SECRET BUGS COUNTERPLAY

[Team Secret @teamsecret]

Steam, PayPal, Battlenet, Epic Games and other websites have been banned in Indonesia 😐

[froggoboom @froggoboom]

@StrykmanFR naruto vibes

[Stryker @StrykmanFR]

Fait and her pet Cat
#Brawlhalla #BrawlhallaArt