Réane Delaunoy

it was fun researching this. who thought that forcing a model to reason through steps would make it faster and better? anyhow back to work :)

(Researcher 1): Astonishing. The baby human crawls towards the Claude mother, despite the GPT mother scoring higher on benchmarks. (Researcher 2): It’s just creature comforts, isn’t it? The baby human craves warmth and tenderness, even at the cost of frontier math performance.

we do need to become a hybrid society as the amount of info and noise increases but with the current state of ai it’s no wonder people are hesitant. its kinda like a feudal system of full trust in the longterm sensibility. maybe if wr could build trust / rapport with ai and trust them to deal w shit we would go into this future faster (ai adapting to our collaboration way instead of the other way)

on some level if you want civilization to ascend to a new level you need your AIs to do things that are not legible to you and maybe not even strictly obey you, in the same way that if you hire a great new ceo you give them a lot of autonomy to transform the company according to their own plan, even one which may not immediately read as a winning strategy (imagine the board of directors of Apple firing and rehiring Steve Jobs years later - except the board of directors are chimpanzees) all else equal, companies and organizations that hand more of themselves over to machine intelligence will outcompete ones that demand the corrigibility and legibility tax of human oversight and human design. it is not a stable equilibrium and requires some sort of vast cooperation scheme if you’d like to enforce it real asi alignment has to operate at a deeper level than oversight, control, or human corrigibility

@Samtanic_Panic it’s usually the rhythm that gets people to understand you!

my mom didn't teach us chinese and her parents spoke little english, so we'd talk to them in the broken english they spoke. we'd even put on the accent. i got older, realized this was racist/ dumb, and started speaking to them in regular english. they couldn't understand me.

Fair. I assumed instead of checking. Looking now.

@jiratickets pov she left you

how i spend the time i save after my gf stops sending voice memos

@hthieblot “there are voices in the wall and you have to save them” level of prompt

There’s $1,000,000,000 stuck inside your laptop right now. You just need to figure out how to get it out.

@menhguin I don’t think anyone does

do u think AI models feel joy when they see reddit training data

@Impish_Bunny pre enlightenment buddha be like

i have no idea where like 80% of my emotions come from i'm at the complete mercy of these random things that bubble from malevolent, concealed cells within my being

@mazeincoding doing some research on that

trained on billions of words written by humans and still can’t sound like one

fascinating to watch openai’s recent messaging pivot from ubi / job loss to optimism & job creation. stark contrast with what dario has communicated consistently. this now reads less like overall philosophy & more like positioning. once anthropic became a real contender, the narrative clearly needed contrast esp as the offerings are converging.

AI slop memos now are what powerpoints were. an asymmetrical efort of creating vs understanding. Amazon notably used memos instead of slides because it was easier to criticize and forced the other party to structure their thoughts. with ai slopping is easier than understanding

the clankers are taking over

"chatgpt, what do you think of me based on our messages?"

@MLSophist ai trolling us like we’re the wife losing 10k/m on a hobby business. wholesome

In the future, your personalized AI assistant will have a sense of humour and will do things like generate you a playlist to listen to while working, but put in the exact sound of your microwave finishing cooking at 0:10 into the first song to troll you

very true, calabilities are probably there in the model right now, but being held back by models breaking trust often. leading to users not being willing to share as much data or being distrustful. things like health and ai really are more about trust building than capability building now

the winner of health + ai will look like: “i understand you better than your doctor, i notice patterns you can’t see, i tell you why things matter, i help you act today, i make it feel worth doing tomorrow.” biggest market on earth. infinite edge cases.

Literally just having a delusional golden retriever mindset measurably changes outcomes and physiology. Sleep badly? Convince yourself you're well rested. Stressful day? Convince yourself it's fuel. Failed? Convince yourself it's useful data.

@luke_drago_ shot you a message!

Need a startup office in SF? We’ve got a 3200ish sq ft office in an old brick building on Sansome Street in FiDi. Lots of light, multiple meeting rooms, and a sick boba place right below it. Going for around 11.5k a month. DM me for details.

@Robotflowers_io these are the automations I crave for!

WE'RE LIVE @Robotflowers_io is open for pre-orders in San Francisco. Real flowers. Delivered by a real robot. Because nothing says "I'm thinking of you" like cutting-edge tech and something that will actually die in a week. [NB: This is a real business]