Pinned Tweet
シカゴで開催されたアニメコンベンションで、盛大なパーティーを締めくくりました。
もう一度やろう!
@S2TB_korsk
dj heroic.vers 「ミッキー」 @ 家 retweeted
Over the past few months, Anthropic scientists have engaged in deep research partnerships with six organizations developing new medicines. We've also begun drug discovery efforts of our own.
In doing so, we’ve learned much more about how Claude can support the many stages of drug development.
In this blog post, we share some early results from these efforts, and describe how this work fits in with our wider ambitions in the life sciences.
anthropic.com/research/early…
dj heroic.vers 「ミッキー」 @ 家 retweeted
People from all walks of life, many far more vulnerable and less fortunate than you, will eventually face the same fate. And I believe that, as some of the earliest casualties of AI, we have an almost sacred responsibility to soften the blow for those who come after us.
For generations, we've tied our sense of purpose and self-worth to what we do, what we earn, and how good we are at it. AI is about to challenge that on an unprecedented scale. We must develop strategies and frameworks for finding joy, purpose, and meaning in things that have nothing to do with economic value.
But there's an extraordinary opportunity here, too. A once-in-a-millennium chance to steer society toward something more caring, more social, more humane. A world where human worth isn't measured by economic usefulness.
That future won't happen by itself. If we don't help shape it, the world your boy grows up in will be shaped by big corpo and utterly clueless politicians instead.
So start with what you already have: your skills.
Remember the response to Cards for Ukraine, heisse-preise.io, and similar projects? How did it feel to build something that genuinely mattered to people? Nobody gave a flying fuck about your tech stack, your frameworks, or how clever your code was. What mattered was that you cared enough to build something that helped.
I'm working on a few charitable projects myself, and seeing people genuinely grateful for something I've built has done more for my sense of purpose than any professional achievement lately. Turns out, being valuable to others is very different from being valuable to the economy, your professional peers, or even your own ego.
And please, get out of the programmer bubble. Touch grass. Hike a mountain. Invite your normie friends over. Talk about anything but AI. There's an entire world out there that couldn't care less about the latest model release.
We're all in this together, but we can't let AI consume every corner of our existence. Especially when the whole point is to rediscover how much more there is to life than the work we happen to be good at.
A guy I greatly respected long before he became one of the rising stars of the AI scene once told me something that's stuck with me ever since:
Be the change you want to see in the world.
I can't think of a better time to take that seriously.
Thanks for coming to my TED talk.
lol this is a good signal on alignment of humanity.
Never let Claude run wild on the privacy policy or legal pages on your website.
Opus 5.5, and other Claude models in general, are so knee deep into moral policing it can genuinely land you in hot water.
> it will get way too specific about how your work happens internally
> it will add promises and guarantees you didn’t want enforced
> it went as far as adding my name in the policy page
> it listed out all the third party tools that we have integrated too, basically showing our entire stack
It is built to write from the customer side by default, it said so when I pressed it. Just be careful.
Man I miss my electrical engineer science fiction loving father.
He would have loved and feared where we are today.
dj heroic.vers 「ミッキー」 @ 家 retweeted
bro that's just vulnerabilityslop. you only like it because it allowed you to open your heart to others
Again. This is fundamentally an alignment issue.
These are problems that we solve via alignment.
There is a disconnect between:
AI is bad vibe coded homebrew
AI poses a threat to all of cyber security because it can find the flaws and holes better than humans can.
Why would AI make bad vibe coded homebrew when it objectively knows how to avoid the flaws and holes?
Replying to @Catruption
well aside from the basic reasons why ai is bad vibecoded homebrew is often unreliable and low quality, and ai now learning how to reverse engineer stuff poses threat to basically all cyber security because it can notice vulnerabilities faster than any group of humans can
dj heroic.vers 「ミッキー」 @ 家 retweeted
just saw You Can See Everything and now this makes a lot of sense
I’m now up all night doing research the same way I did in high school and college with nonstop Wikipedia.
But I’m now 35 and the research gives me anxiety and then now I have anxiety induced diarrhea.
No more screens in bed. For real.
Hahah SEGA isn’t using AI to fix their AI vulnerabilities and now Aime is restricted to just Japan lol.
Talk about a cheap bandaid fix.
【Aimeサービスサイト、AimePayメンテナンス終了】
10/21(水) 22:00 より実施しておりましたメンテは、14:00 をもって終了いたしました。
メンテへのご協力ありがとうございました。
詳細は下記お知らせをご確認ください。
my-aime.net/news/detail?news…
Normies should know about this because it will actually help them understand what’s going on here.
Now if they can handle the truth? Well. That’s another question.
dj heroic.vers 「ミッキー」 @ 家 retweeted
normies *don't* know about this
normies don't talk or know much about ai at all
if you push real hard you might get some hand-wavey "datacenters are bad because uhhhhhhhhhhhh they just are" sort of response
they're just trying to pay rent for the month, twitter is a bubble
Hey has anyone considered the implications of the software revolution occurring at the same time as the midterms?
Like, voting machines should assumed to be compromised, yeah?
Yeah, if you want to choose that.
Or, all of humanity got a giant level up in cognition and you can now be more inspired than ever to continue to learn.
This is how I realized that memory management may as well be another layer of the agentic IAM onion.
No harness today has an off the shelf good way to keep clear boundaries between memories.
Claude projects are a step there but if project an automatically sees everything…



























