Wires

the future is infra

[Wires @internetiswired]

@finalquestline @jallen_dev curiously tho are you really even learning anything if ur just prompting the ay eye and asking it how to fix?

I guess the nuance is just that they need to engage with the error rather than "just fix it" to actually learn something including reading the code

[Wires @internetiswired]

@l33telite1337 @vxunderground im talking about the guy being quoted not vxunderground lol

[Wires @internetiswired]

one shotting apps has never been a good benchmark if you haven't figured it out yet

[Wires @internetiswired]

babe it wasn't me it was codex, see i even tweeted about it

[skooks @skooookum]

Just saw a pornhub URL in the Codex thinking trace while I asked it to write me a small shader. Whatever man do your thing

[Wires @internetiswired]

1) what

[Prime Intellect @PrimeIntellect]

Prime Agent is a general-purpose coding harness

On ARC-AGI-3, it scores 95.5%, surpassing the human-expert baseline, but the gain is not benchmark-specific.

We see major improvements across models when compared to their proprietary harnesses:

[Wires @internetiswired]

@zanyfen i feel like many people don't understand there's nuance to ai usage. Goods and Bads, it's a good search engine that can cite sources. I can click the source and confirm my answer + find even more information thats needed on the cited source

[Wires @internetiswired]

@kennethnym this twitter account unironically bankruptted me after i watched a twitch stream and heard his keyboard ONCE

im on my 3rd hhkb

[Wires @internetiswired]

@eisokant i am excited to see the future of where your models go

[Wires @internetiswired]

the future of compute is in infra and purpose built models that you can load at a glance, i believe this

[Wires @internetiswired]

haven't gotten a chance to demo this out but excited for the future of highly purpose oriented models

[Poolside @poolsideai]

Today we're releasing Laguna S 2.1, our most capable model to date.

It's a 118B total parameter Mixture-of-Experts model with 8B activated per token, a context window of up to 1M tokens, and thinking and no-thinking modes.

Capable enough to hold its own against models many times its size. Small enough to run on a single @NVIDIAAI DGX Spark.

Laguna S 2.1 is fully open under OpenMDW-1.1, with weights available today on @huggingface

https://t.co/xxGeAgo35R

[Wires @internetiswired]

rebuilt my portfolio with an obsession for the details :), design is a signal

adapt and conquer

[Wires @internetiswired]

i need to acquire more compute

[Wires @internetiswired]

im back and im gonna tweet and interact about ai stuff since i work there thanks

[Wires @internetiswired]

@Dave_Geoghegan_ @shydev69 fully local but requires a gemini api key?

[Wires @internetiswired]

kinda crazy how much i think schools affected my actual learning in life

[Wires @internetiswired]

i noticed i really enjoy learning when im not constricted by forced documents or strict deadlines

[Wires @internetiswired]

the odin project has been pretty fun learning

[Wires @internetiswired]

i wonder how nuts id be if i wasnt limited by mmo games hehe

[CRONK 🎩 Crypto Reporter @CryptoCronkite]

literally crying rn

my pregnant wife went into labour today but I had to stay at my computer for the BAYC land mint so I missed my son's birth and I didn't even have enough ETH for gas

fuck crypto