technological progressive. voted most trustworthy tech news source 1898. π§¬
Colorado, USA
Joined February 2025
- Tweets3.2K
- Following789
- Followers1.2K
- Likes78.6K
Pinned Tweet
iβm only really just now updating to 3sum being subquadratic and i think the reason itβs shaken me so badly is it proves that algorithmic space is densely populated with alien solutions that our brains are fundamentally ill-equipped to find. thereβs a twelve foot tall man in the corner of our house that none of us can see.
thereβs a gem hidden deep in this thread if you dare to search
whenever i see someone has me blocked i instantly block them back in sort of a scorched earth policy
1
25
434
sort of funny seeing a brigade of people in the replies saying this is baseless and just bad prose and then you look at the copy and itβs the most straight-from-claude shit youβve ever read
queens domain is the only game since dark souls that understands what made dark souls great. itβs not even a βsoulslikeβ!
fable 5.2 is a just ok model that i think makes it clear that the current paradigm of training these things on the traces of web developers is bad because web developers arenβt really people
really, really interesting thread. reading some of the examples triggered fight or flight
Why is AI slop so frustrating to read?
AI writing often obscures main ideas and invents awkward phrases, making sentences more complex than necessary.
Many of these slop patterns violate writing principles established in English writing books.
In our benchmark, AI writing breaks these principles far more often than human writing, with Claude Code replies by far the worst offenders.
We show examples of the most common slop patterns in this thread
ALT How often AI and humans break writing rules
a very underrated model, i think. the only one of the whole deluge of system one models that's been interesting (since jev jevself)
a brand new model paradigm! introducing πππππππ: a system one model for extracting, computing, estimating, and predicting values from unstructured inputs.
blog.cloudflare.com/clairic-β¦
pretty cool to see a company publish charts where their models aren't sota. what the hell is opus doing
Introducing Inkling 2 Small. Scoring 31 on the AA Intelligence Index, it's a big jump over its predecessor.
Inkling 2 Small is available for fine-tuning and inference on Tinker today, with full weight release planned soon.
thinkingmachines.ai/news/inkβ¦
a very interesting model with a very unique style!
Today we release Ripple-65B Preview, our 65B dense LLM we pretrained from scratch.
Ripple sets a new state-of-the-art for U.S. open-weight models in agentic coding tasks, surpassing significantly larger models including Muse Spark Open-816B, Beam-501B, Inkling-975B, and Nemotron 3 Ultra-550B.
Learn more: ricursive.com/ripple