Andy Walters

@mikedotexe in descending order of difficulty for AI to cannibalize: taste, accountability, judgment, agency

recent AI models like Fable and Sol are so capable, i think humans should call for a worldwide symposium where we talk about "what human work is good for anymore" for millennia, the human race operated under scarcity constraints that no longer apply

The doomers were wrong, as anyone outside the fear bubble with a cursory knowledge of history could see. Technology does not instantly diffuse through an economy. And every technological revolution reveals value in human abilities we previously took for granted. In the age of AI, those are judgment, taste, accountability, and — yes — agency.

@willdepue i love this… we’re building this for Enterprise AE’s. they’re gonna fucking love it.

dear claude code & codex teams, please, for the love of god, where is my executive super assistant that has: (1) a deep understanding of me via great memory, just pack 200k context with every chat. you can build this personal store from past chats, but also i'll just give you all my data, respond to 100 different personal questions, give you all my Apple Notes and iMessage (2) a no-chat interface. i don't want something that forgets me everytime, that i have to skip to the right chat. just ditch chats! its all one chat history, same as a real assistant. compaction and retrieval cannot be that hard (3) i can trust to reasonable certainty with my credit cards, info, email access, etc. and i can delegate to go do things for me. i just want something that can book a flight, hotel, refill my runpod credits, update all my old passwords, cancel all unused subscriptions (4) it is good at pinging me by sending me a text when i need to answer something for it or do something it can't. good back and forth communication. it also can't die all the time. i do think the idea of 'responses' is wrong, it should just always be 'on' in idle mode, ready to act on scheduled tasks. it should have lots of webhooks/awaits, for example, maybe it should text me everytime i receive an important email that needs my attention. a small model can scan everything and then call a tool to wake up BIGGER ASKS: (5) please let it call people for me. go schedule drs. appointments, sit on hold for me, etc. same with access to imessage/email etc on just handling things for me that i don't want to care about. (6) computer use that doesn't interrupt my work (works on another desktop on my mac/its own windows?) and is faster than me at filling forms and doing things. like how are models slower than me at typing and filling out forms than me still! i will pay you $5k a month minimum for this. if you need me to buy some physical hardware or rent some cloud thing that's fine. i'm fine with taking on some security risk, it's that valuable, you take risk with a human EA too chatgpt almost feels deprecated as a product compared to codex now. like models in codex are significantly less helpless and more capable than chatgpt, such that i don't want to use chat anymore (other than as a google replacement). plz fix or just build an integrated experience - will

@SusieWiles47 no i am not thanking you for delaying what should have been a perfectly routine software release no your priority is not the fastest deployment of the tech, it is security theater

Under President Trump’s leadership the United States is the undisputed winner in the AI race. My gratitude to companies across industries who continue to work closely with the White House to implement the President’s EO: “Promoting Advanced AI Innovation and Security.” This includes excellent work around advanced model access and guardrail testing and security. The government and private sector have worked together in a way we have never seen before and this foundation of America First is unprecedented. Our shared priority remains: get the best tech deployed as quickly and safely as possible.

i cannot begin to describe how annoyed i am that i am rooting for anthropic right now

there continues to be a shocking amount of alpha in simply rebuilding workflows to be ai native the best startups can’t help but reinvent their market because they just get annoyed at having to do things the old way then imagine how few teams are doing this at any given time: incumbents are handcuffed to their business model and even in supposedly large markets there are only a handful of teams with the funding and workflow knowledge so you end up with a situation where if you can cobble together a little funding a half decent vision and a little distribution magic, you have a serious shot at winning

@petergostev this was so obviously intelligible i didn’t realize it was satire until i scrolled two tweets past it haha

4.6 was my daily driver, but 4.7 and 4.8 didn’t do it for me, hard to go back after Fable got blocked. Using 5.5 a lot more now, can't believe I can't use 5.6. 5.2 was impressive, can really see myself switching. No idea what’s going on with 3.5, it’s been months since 3.1.

@FDatframe84228 honestly i’m not confident a biden/harris administration would have handled this better

@andywalters Or maybe Trump should have not sold “no regulations” BS in the begging of the year and left the ones that actually made sense in an actual framework from Biden.

well they won. for now anyway. the bay area rationalist yudkowskiites sowed enough fear and uncertainty into the discourse they created the dystopian hell we now inhabit in which Trump has to personally approve every model release from here on out. truly idiotic

@latentlocal *epistemic

while i am no fan of this administration, their decision did not occur in an epidemic vacuum. do you think we would have been here if the alarmists hadn’t spent years making wildly inflated, unsubstantiated claims about the harms these models can supposedly inflict? and advocating for draconian and utterly disproportionate responses such as bombing data centers?

i was never under the illusion that trump was anything but an autocrat who wantonly abuses government power i am not in principle opposed to regulation if demonstrable harms to the common good are at stake, regulation is transparent, fair, and swift, and regulation does not materially disadvantage the national interest. it is unclear to me that any of those three conditions have been met.

@andywalters Trump was always going to seize control and use it to extract concessions/ensure loyalty. You were fooling yourself if you thought he'd be libertarian Yall act like there's no difference between eg Russia and Finland. That all government action is equally currupt/ad hoc/bad

@lightscaner who do you think lutnick is taking orders from?

@andywalters It was howard lutnick who sent them the letter. You think Trump is running the show? He's just the frontman / fall guy

@_coenen hard disagree but i can respect that it’s an intellectually coherent position

@andywalters counterpoint is that things are moving too quickly right now - overdeployment of capital and AI job disruption are both massive economic risks. and there’s tons of headroom even in pre-fable models. might be very good long-term to cool things down a bit

@UgurOzk34988011 Access to intelligence is the new human rights issue

@andywalters Govs and certain conglomerates: frontier ai We peasants: retarded version with different names

@robleclerc @kevinnbass They nerfed bio so hard I saw a screenshot of someone asking if mitochondria was the powerhouse of the cell and it refused to answer haha. Utterly absurd

@andywalters @kevinnbass they were easily able to nerf bio topics, would it have been that hard for them to nerf cyber-related prompts as well?

@Nymne Agree

@andywalters I know that people want us to blame Trump for this, but that’s like blaming the bull for attacking after waving a red cloth in front of him. In this case the red cloth was the personal power of regulating a whole industry. Of course Trump was going to take that opportunity

@jxnlco never going to use anything but xhigh until usage limits force me to think about it

when you use codex do you want to choose the model and reasoning effort yourself? should the product just pick based on the task? what would make you trust that choice? Are you someone who always uses extra high, or do you think about using low when you don't need that much thinking

@petrusenko_max If you’re looking shoot us an email — apply@agent.press

@andywalters Why do I see startup roles. Brrr

Hiring a full stack engineer! Apply if you love building, love AI, and are up for the startup ride 📈🚀 agent.press/careers/full-s…

@HarrierOnChain Interested — Shoot us an email apply@agent.press & mention I responded to you on X

@andywalters Built a couple full-stack AI agent platforms last quarter — Next.js frontend with tool-calling backends. Fits the startup pace well. Happy to share a relevant repo. DM?