π¨π»βπ» Principal Software Engineer && π» Front-end Chapter Lead && π Designer && π Design System Lover ESFP-T
nulldefined
Joined November 2007
- Tweets26.6K
- Following877
- Followers2.5K
- Likes12.5K
Nass.js π retweeted
most people are using AI to build more.
β more features
β more pages
β more stuff
i believe the bigger unlock is using it to make one thing really, really good.
β speed up your slowest endpoint
β make CI faster for the whole team
β fix the bug everyone stopped noticing
Nass.js π retweeted
VΓdeo 1: uma excelente explicação sobre o por que a Bolsa caiu: a aposta dos investidores foi em FlΓ‘vio e deu errado. VΓdeo 2: um exemplo prΓ‘tico do que o vΓdeo 1 explicou. Vejam um investidor desesperado pela derrota da aposta que ele fez.
Nass.js π retweeted
Holy sh*t! Opus 5.5 (Launch) vs Opus 5.5 (Now)
Same prompt, same effort.
The usage: They nerfed the usage limits. Hard limits are draining MUCH faster !!!
The quality: It feels similar no big change but sometimes you feel a little laziness but other than that, it's the same
Anthropic, why do you nerf the limits?
Nass.js π retweeted
IBOVESPA during Lulaβs previous terms:
Term 1: 9,000 -> 65,000
Term 2 (pre-2008 crash): 65,000 -> 73,000
*crashed from 73,000 down to 30,000 during 2008
Term 2 (post-crash): 30,000 -> 70,000
Term 2 (entire term): 65,000 -> 70,000
Term 3: 114,000 -> 180,000
~23% per year (1100% cumulative) for the 12 years in office, including 2008 crash
~4% per year (63% total) for the 12 years he wasnβt in office from 2010-2022
index currently down ~25% overnight
Nass.js π retweeted
RIP em dash
We analyzed how @claudeaiβs Opus 5.5 writes compared with Opus 5 across high-reasoning Text Arena outputs.
10 of 12 writing measures moved in a better direction.
Opus 5.5 should be easier to read:
- Long content words fall from 41.7% to 38.6%, the lowest share of any Claude model we analyzed.
- Sentences are 17% shorter on average, dropping from 12.14 to 10.03 words.
The tradeoff is length. Answers get 6% wordier, rising from 453 to 481 words on average, making Opus 5.5 give the longest answers across the Opus family.
It also sounds less recognizably AI on two familiar tells:
- 95% fewer em dashes
- 73% fewer semicolons
But a new giveaway may be emerging. Hedges and caveats such as βperhapsβ and βarguablyβ rise 97%, from 0.39 to 0.77 per 1,000 words, the highest rate of any Claude model we analyzed.
What do you think: does Opus 5.5 read more naturally?
Nass.js π retweeted
δ»ζ΅θ‘γ£γ¦γγγ»γΌγ©γΌγ γΌγ³ε€θΊ«εη»
ε―ζγγγπ€£π
Nass.js π retweeted
We can finally talk about it:
We found a way to extract hidden reasoning of frontier models using a vulnerability in the APIs of every frontier AI company.
We verified that our reasoning token count matches billed API thinking tokens 1:1 for most of the prompts we queried.
