Rogerio Hirooka

Scientia est lux lucis - iOS Developer, working at SkipTheDishes - Founder of LirumLabs - Brazilian born Japanese living in Canada 🇨🇦 🇯🇵 🇧🇷

[Rogerio Hirooka @rogerioth]

@manitobahydro #mbstorm the power outage is 12+ hours already in Transcona / Mission Gardens…

[Rogerio Hirooka @rogerioth]

@XH_Lee23 She forgot about the 5G ban from years ago, the drone (DJI) ban from a few months ago…
Less and less freedom every week

[Anonymous @YourAnonNews]

These anti-science fucks are the same assholes from the 1600s and 1700s who tied women to logs and set them on fire because of 'witchcraft'.
When stupid people are in charge, witch hunts always happen.

[Rogerio Hirooka @rogerioth]

#mbstorm @manitobahydro
Received an email from Hydro stating that the power is expected to restore at 3AM

🤞

[Rogerio Hirooka @rogerioth]

@ShelbysFade @MissSophiaaxx Some fans forget that what makes GoT products so good is the realism in fantasy

[𝘊𝘰𝘳𝘳𝘪𝘯𝘦 @OopsGuess]

You mean Huawei “stole” the 5G America never had?

And now Chinese open-source AI supposedly “distilled” its way into beating America’s top frontier models?

Amazing.

America keeps accusing China of stealing things it cannot build, cannot price, cannot open-source, and cannot compete with.

That is not an IP case,
it's simply that the United States cannot accept defeat.

Again and again.

If Washington cannot accept losing, just say so.
If it is desperate enough to use sanctions to stop China’s progress, be brave and be honest.

There is no need to dress insecurity up as intellectual property law.

And since Washington suddenly loves talking about stolen data, let’s start with OpenAI and Anthropic’s licensed training data.

Show the receipts.

[Rogerio Hirooka @rogerioth]

@caiocgomes Os que ‘importam’ pra eles infelizmente são bastante idiotas… e racistas, e todos os adjetivos…

[dax @thdxr]

"One probable outcome of an open-weight-model-dominant world is full AI communism"

[Rogerio Hirooka @rogerioth]

@PatrickMoorhead Dude, that's not how distillation works...

[Haider. @haider1]

PREDICTION

after kimi k3 successful launch and mogs opus 4.8, Dario Amodei is gonna write an essay explaining how open-weight Chinese AI threatens democracy, national security, and the future of humanity

[🐝🇬🇷 @bee_fumo]

history repeating itself

https://t.co/qvqaFBQ4ok

[Rogerio Hirooka @rogerioth]

@kimmonismus He is competing to become the most disgusting CEO of them all.... and it's a harsh competition

[Derya Unutmaz, MD @DeryaTR_]

DeepSeek should ban Anthropic from implementing Dspark into their models! In fact, Anthropic should be banned from using any AI research from China and required to remove any non-US data they used to pre-train their models! Cry, Dario, cry!😅

[Rogerio Hirooka @rogerioth]

@TheEconomist What’s your definition of ‘efficient’?
It uses less tokens to achieve the same result;
It uses less energy, take less compute to train;
Take less time;
Costs less for the inference;
… what is “more efficient” then?

[Conor Dart @Conor_D_Dart]

@ChaseWisely I did give that some thought, but when opus is doing really stupid crap that even sonnet was doing fine then i know something is up for sure

[Forrest Knight @ForrestPKnight]

Opus 4.8 has been super dumb today. Just me?

[Dan Kulkov @DanKulkov]

opus 4.8 is back to 40 IQ retard who needs to be micro-managed :(

[Crow @CursiveCrow]

@ValsAI an incomplete task is a 0, not "do it again with a different model"

[Shannon Sands @max_paperclips]

@ValsAI so, you didn't actually score Fable then, you scored Opus. LMAO

yeah, this ain't it

[Conor @jconorgrogan]

@ValsAI You should absolutely zero out those scores. A refusal is worse than a half-hearted effort with bugs. I dont see why you rated it this way