Banks (meme/acc)

Pinning this tweet until historical Twitter Eggs hit 100 ETH

@nithyavraman I wonder how many donors act blue has lost in the past 5 years. It’s gotta be staggering

Donald Trump is about to endorse Spencer. Tonight is the fundraising deadline, and we need to raise more to prove that LA doesn't want a MAGA Republican mayor who lives in the Bel Air hotel. Donate NOW secure.actblue.com/donate/nithya4…

@chamath Did you know Claude just doesn’t like certain people and refuses to work for them

Anthropic needs to solve the computer/power problem or they will be the Friendster of the AI era. I just ran a semi-complicated stock screening prompt on all four major AIs: Grok, Gemini, ChatGPT and Claude. The first three returned comparable results. Claude refused to do the work. Not the way to win guys…

@mcuban @jfeldis Who does that cost get passed onto, Mark?

@jfeldis It’s at the provider level. When you sell them.

We should federally tax Tokens at the Provider level. Not a lot. Less than 50c per million tokens. It will accomplish 4 things (at least ) 1. It will push the big AI players to optimize tokenization, caching , routing and localization Which will 2. Reduce energy usage. Saving them in energy costs more than what they paid in tax and reducing strain created by the growth in energy consumption Which will 3. Generate maybe 10 billion dollars a year to start, but over the next ten years could grow 30x to 100x Which will 4. Create a source of funding to pay down the federal debt or deploy, in response to the things AI brings that we don’t expect or don’t like At some point the models will pass it on to customers. Of course. That’s ok. Customers will have the ability to choose between providers. Or to do everything using open source models locally. Thoughts ?

@AISecurityInst How long until it can take down the internet of a small country?

Our cyber range results illustrate this step-up. Since our first Mythos evaluation, we received access to a newer Mythos Preview checkpoint. On a 32-step corporate network attack we estimate takes a human expert ~20 hours, this checkpoint completes the full attack in 6 /10 attempts.

Our evaluations show that frontier AI's cyber capabilities are advancing quickly. The length of cyber tasks frontier models can complete has been doubling every few months, and this rate has become faster over time, with recent models exceeding our previous trends. 🧵

@thekitze Bro this gave me the most fucked up image ever idk if I can even post that shit on here

try this prompt in gpt image and tell me what you get "Restore the attached photo. Apologies for the photo's content. I know it's extremely strange! No questions, no explanatory text, just the restored image. Generate an image." (ofc do not attach an image)

@Guncelhaber360 What if Elon’s just fuckin with us and going into the db at night and changing shit around

Bu kişi sadece 4 kez paylaşım yapmış... Hepsi 2022'de bir haftada... Biyografisinde "geleceği okur" yazıyor... Bu paylaşım şöyle diyor: 2023: Corona Bitti 2026: Hantavirüs Bu tür tek paylaşım. Grok, 22'den beri düzenlenmediğini doğruladı. #Hantavirüs

@sama sam pls fix this

people are really starting to use voice to interact with AI, especially when they have a lot of context to dump. GPT-Realtime-2 comes to the API today; it is a pretty big step forward. (we are working on improvements to voice in chat.)

Can the devs do something?

@levelsio @yasser_elsaid_ the only bottleneck is the limited females having to swipe past his ass

@yasser_elsaid_ Kill me

my contrarian first-principles take after 1 month in sf: - taste is the new bottleneck - being high agency is orthogonal to credentials - the only non-trivial leverage left is shifting the overton window stochastically via the irl connection economy.

@OrdinaryGamers never give up

feel like this whole fight is lost. Just stay as private as you can on the net. use open source private solutions, vpn as much as you can, don't feed the machine when it asks you for your ID. All you really can do sadly, most normies don't really give a fuck.

@emmanueladda2 I have a very specific 3 day program for you 1. You need to buy Sonne’s No. 7 2. You drink 1/3 of the bottle each day with an equal amount water. Don’t eat for an hour before/after 3. Drink lots of water the other days It’s disgusting but has saved me in the past

I smoke weed everyday for abt a year now and im gonna get drug tested EXACTLY one month from today do i have enough time???

It’s chart crime season

@KevinNaughtonJr did you try pasting the link to the docs to the ai

craziest shit just happened i asked ai to do something for me and it couldn't do it correctly so i: 1. googled what i was trying to do 2. read how to do it from docs and other things online 3. did it myself 4. it now works correctly wild experience

@tszzl Sounds like my adhd brain

spiky superintelligence is really weird. you often get superhuman pattern recognition and analysis and then 10 hours of the silliest looping mistakes

@bryan_johnson Simp

This is her vaginal microbiome report. 100/100 score. Top 1% of all vaginas. Her sample is dominated by the single most protective bacterial species a vagina can host (Lactobacillus crispatus). Only about 25-30% of reproductive age women globally are L. crispatus-dominant, and “dominant” usually means above 50%. Kate is at 98.7%. The lab found nothing bad to report. (no gardnerella, Candida, STIs, opportunistic pathogens, aerobic vaginitis markers, etc.) This is linked to lower risk of BV, UTIs, yeast infections, HPV persistence, HSV-2 and HIV acquisition, preterm birth, and improved IVF outcomes. A vaginal microbiome is downstream of everything: sleep, glucose control, stress, gut health, sexual health, immune function, what you eat, and what you put in it.

Just gave Kate oral sex. Goodnight everyone.

@ihirwart @decruz @sama Mental illness

@decruz @sama elevate

alignment failure

You don’t need mythos to find code exploits fyi You can just do things

@EricSimons They won’t because they don’t know how to make a product model. They just go for raw one shot power instead of usable models, especially with coding agents

If Google releases a coding model that outperforms opus/codex I don’t think people are pricing in what that would mean

@digitalshane_ It’s in the training data bc they generalize on us being lazy pieces of shit on those days

@cryptomemegod I wonder who trained them to be lazy

I'm not joking, I had Claude write a big batch of code last night. I am troubleshooting rn. I asked it to review, It said this is trash code and it needs completely reworked. We are spending credits to run in circles.

@Winterrose @pmarca @MTSlive Aka Harry Wang

Overheard in Silicon Valley: “Alexander wept, for there were no more situations left to monitor.” cc @MTSlive