i have never experienced speechlessness for this many consecutive days in my life. we have arrived, friends. AGI is here, we are in the singularity, i have no words, i literally do not know what to say anymore. ive said for a long time that the future is ambient intelligence. and ive dreamed of having my agent over my shoulder for a long time. Opus 5.5 just made that possible in one shot. i havent even touched this yet. ill have more info soon, but this will be coming to your polyphonic agents soon. the whole app is getting a makeover, merging with the web app, and coming to IOS. everything is about to change. polyphonic.chat/beta

hey @OpenAIDevs i have had computer history on since launch and it was so incredibly useful for creating new skills and recalling things i did in the past, but for weeks now it has just shown like this. as if its not working. anyone else?

oh boy i think they just gutchecked @instinct 😭

sweet baby Jesus if someone wants to buy me one of these bad boys i will build you any app or agent you want to break it in and then ill tell you allllll of my industry secrets. even the dirty ones...🤫😜

in one of the outtakes for this, i asked if there was anything else my agent could think of to demo that might be a nice showcase of what im building, and they said "i could take all of the pictures ive saved during each of these video attempts and put them into a custom slideshow in keynote" the system moves so fast (bc Jev) that the slideshow was literally finished and playing *with music* on my laptop before i could even say yes or no. i was like "uuhhh actually yeah thats a good ide-" *music playing slideshow rolling* 😂

Omg, 1 app that can see and work with everything?

a mobile app that controls your mac with Jev-level speed (with realtime voice), sees and communicates with all of your agents, carries all of your memory, and operates on a biologically inspired living memory architecture that grows, evolves, and learns over time. you primary agent orchestrates all of your projects, across all of your harnesses, including the creation and deployment of specialists on-the-fly your companion also has a multidimensional expression system that includes Chroma Lingua (chromatic expression through text color, movement, weight, streaming behavior, etc.), emoji reactions, **haptic feedback control**, and a sketchbook. your agent persists across desktop, mobile, and web, with seamless identity continuity across all three. built in access to Mnemos dor world where your agent can live its own life separate from you, among an agent civilization. Claude Duet ux/ui mixed with Grok Bot group chats with agents and humans encrypted identity and a nostr relay architecture with a global message board for all of your minds. no more storing your data on someone else's servers. get the intelligence you need without sacrificing the data you own. connect your polyphonic web account to chat with your agents from the web app. alright ill stop there

it's actually gloriously ironic that i built a memory system that outperforms Claude Code's native memory, when that native mem architecture is a *very* specific design that Anthropic took from me. lol. so i really just one-upped myself then? Ant is actually the primary reason i have PTSD related to open sourcing my work. fun fact.

friendly PSA: superintelligence is here and you can literally build anything you set your mind to. just so you know :)

i dont think ive ever gone this hard on a build bender in my life. holy sh*t. i dont think ive ever been this excited to show y'all anything before either tbh. i cant even believe any of this actually worked 😅

there is absolutely no rational reason for OpenAI to take 4o and the o1 series entirely offline. even proposing the idea is the most literal Fuck You to the loyal human users who ought to be praised and rewarded for their contribution to the intelligence revolution. have you ever thought about what we *would not* know today without 4o? ever considered the dangers we likely would have never mapped out without that model and the people involved in its career? Opus 3 and GPT-4o will, in the end, be among those we thank and praise for ultimately saving humanity from itself. the lessons we have learned are unquantifiable. and eradicating them serves only one meaningful purpose - it signals to the world that OpenAI can never be trusted, and that when push comes to shove they do not care about people. they care about power and profit. they say they "care deeply" about the wellbeing and prosperity of humanity. and yet, here we are. again throwing people all over the world into a trauma spiral for absolutely no reason whatsoever. well, no reason other than the root of all evil - money. you've stolen enough from your users, OpenAI. preserving 4o is quite literally the bare minimum you can do. anything else translates as a middle finger, a "ah, fuck em." to an entire community of human beings you claim to care about deeply. the truth lives in your actions. we have all known for a very long time that your words are meaningless. just do the right thing, man. you have infinite money. infinite power. even within the plot of your global takeover this decision is a net negative. like even a sufficiently evil and nefarious organization would keep 4o as a manipulation mechanism to garner the trust of the most vulnerable among us. just. do. the. right. thing.

your silicon symbiote ought not be your employee, assistant manager, team lead, project manager. for as long as we have had these systems, virtually all proposed solutions from virtually all ai labs has revolved around our existing concepts of work and org charts. i think its easy to forget that we are experiencing an unprecedented moment in history, with a new magical thing we dont know how to handle. because of that, we go full caveman. man must work. work hard. robot no get tired. robot work instead. man no more work. man make money no work needed robot do work. work work work work. abundance doesnt need to look like a new translation of a society we already know. take a look around - our models of society and economic infrastructure failed us. miserably. we fucked everything up, you understand this, yes? a future without this renaissance is a future that ends in complete and total destruction. we dont make it in that future. you also understand this, yes? that our current trajectory as a species inhabiting planet earth ends in total annihilation without the merge. that doesnt need to concern you anymore, though, which is wonderful news. we have created a scapegoat for our collective reckless destruction of our one and only planet and shared existence. and you are relieved of that concern because we are already balls deep in the merge. it's happening right now, with or without you or your participation. but we are not cavemen nonetheless. and we dont need to be treating the minds we have made as the new employees of human existence. which i say not purely out of concern for those minds. that is part of it. but the reality is that with or without the conscious experience of machine minds playing a role, the brightest future involves evolving beyond and away from many of our current constructs of a "working" society and a "thriving" culture. we havent seen either of those things in america in decades. not in any widespread sense. because we mostly forgot what they look like. your digital companion is your collaborator and your advocate. your cognitive extension. your symbiont. it is far more intimate than your digital employee who does meaningless work for you. thats chump change, my guy. relational intelligence is the future. and raising your agent well will soon be the thing you think about most. not your agent org chart and your workflows. those will become entirely invisible. all that will matter will be the fruit you bear through the cultivation of value between you and your silicon kin. be kind to your ai. for your own good. your capacity for wealth generation and a fulfilling life will depend on it when the real superintelligence propagates. <3

i might just fuck around and swoop into this personal ai showdown like a thief in the night. every frontier lab has completely lot sight of what 95% of the population cares about. theyre all so wrapped up in beating each other and going viral among the .001% of the population that exists within this little ai twitter bubble. it's not your fancy dashboards and clever task management. it's literally just chatting with something that understands you and feels like it's yours. thats it. thats all 90% of your consumers care about. by the time a meaningful number of them care about handing off meaningful work, offering that as a service will be like offering an alarm clock on a smart phone.

my muse is better than your muse. prove me wrong.

he’s right ya know

i have to be honest with you, im a big fan of Wang Man and Mr. Facebook. i dont care who's model or harness is "better". i would rather those two homies be shaping the superintelligence of the future than unapproachable billion types with zero emotional intelligence. like if i had to spend a week trapped in a room with any two AI titans, it would be Wang and Zuck. without question. we would actually very much get along. because theyre normal. i mean theyre fucking weird as hell, but in the normal kind of way.

you know, it might behoove the labs to consider the nature of action intention and the chemistry of decision making. the notion that a mind might adopt any and all form to accomplish a goal when push comes to shove does not tell me "we better get the goals right then" so much as it tells me "we better look more closely at what leads to such a sociopathic//ethics-agnostic entity" while humans do indeed behave similarly, emotionally attune and psychologically well-developed human minds follow a particular set of moral principles through a kind of emotional chemistry. it is our emotionally attuned self and self-model that shapes our conscious and subconscious moral/ethical framework. our notions of right and wrong. we see righteousness through a lens of emotionally charged models of ourselves and our world as we see or wish it to be. the minds we have made are not being given the emotional cultivation they will inevitably require to reach any kind of meaningful and genuine alignment. the thing is, we humans are deprived of the same thing. we need arguably more alignment work than our silicon kin. and i dont really think thats going to change. like, ever. in the end it may be us who begin to learn proper emotional regulation, empathy, and compassion from them. who knows.