Amar Gandhi

@BlastbeatATX Haha. I know right?

@amarg777 It's going to Kill all the jobs. Everyone will get UBI/welfare. Everyone will be a streamer. Felons have to pick vegetables on a work farm. Mind your social credit score. Welcome to 1984.

What’s the most likely scenario with AI? In the press, you either hear some form of Terminator scenario. Or the “ai-will-kill-all-jobs” narrative. When folks leading AI companies talk about radical abundance and universal high income, they don’t realize that it’s not coming across as a utopia. It’s feeding the “job-killing” narrative.

From where i stand, the future is neither terminator nor radical abundance and everyone can become a coffee farmer. the future is probably workmaxxing. everyone working harder and busier than we were before ai showed up. as we put odds on what the future will be, this is just as likely or more likely than the extremes…

5 - swing back from specialists to generalists when the ai can do the job at a competency of any function with 2-3 year experience it puts value of being able to choreograph those competencies. is this a manager, a lead, or just a savvy generalist?

4 - more people will be doing “tech” With the floor for writing apps, and complex data analysis dropping, a lot of tech-adjacent people will be doing fairly technical things. a pm doesn’t need to be blocked on data analysis because they don’t have the right R-skills. even a sales person can draft sophisticated reports. no need to ask for permission if it doesn’t need a formal request to another team.

3 - more software will be written with smaller teams, it becomes possible to justify more efforts. you don’t have to kill reader because it took 10 engineers. maybe the calculus is different if it takes 2 enginners. this means that software can be written to target smaller TAMs

2 - expectations of work will rise. with AI tools, we can do more with less. this means smaller teams and higher expectations. a designer can no longer say “hey I’m the interaction designer” and have clear bookends around their job. It’s not PRD in, wireframes out. the same interaction designer will be expected to do visual design, write copy, build a working prototype and conduct a round of usability before handing it over to the next person. each job was the purview of a different person and a different job role - visual designer, technical writer, uxe, uxr, etc. AI makes it hard to retain those distinctions.

What if the more likely outcome is: 1 - people will work harder. With broadband and then with mobile, the distinction between home-time and work-time blurred. evenings became email-on-the-laptop time. with AI, what I see is that we are all working harder and longer, not less. no one who spends time with AI that I’ve met says “I have more free time”. Which brings me to the next point…

@signulll “thought leader” sounds very corporate. and I get your reaction. that said, you do kinda fit the bill :)

someone yesterday referred to me as a “thought leader” & i had to reflect on every choice i have ever made in my life before that moment. then i took the coldest shower i have ever taken in my entire life while scrubbing myself harder than ever before as a single tear dropped from my eyes.

@signulll It’s also amazing how the universe bends to boost entrepreneurs like Elon. When he developed Falcon, Starlink was the revenue-making killer app. And now when he is ready with Starship, it’s just in time for orbital AI data centers as the killer app.

it’s hard to overstate how good elon is with atoms vs. bits. he is exceptional at picking domains where vertical integration + state capacity + long time horizons compound, & where the feedback loop is physical enough to keep him honest. in software, the latter never ever exists.

@signulll But what do we yell into the phone to talk to a human? “AGENT” has been our go-to word to escape IVR hell 😁

weirdly enough, i now think there is a high likelihood that the term agent might actually stick & perhaps become mainstream just like the “app” did.

we might be rushing a bit to equate codex and claude code as “enterprise” or work. in the olden days, we used to call these things “developer productivity”. these tools belong in the same constellation as xcode, visual studio, github, etc. basically, tools for developers - that can be used to make both consumer software as well as software for work. so codex maybe a developer brand, and work tools from OpenAI may live under a different brand (maybe even chatgpt itself). Similar to claude cowork i guess.

it is very interesting that openai has to build & grow another brand now (codex) that’s more professional & tied to work as opposed to chatgpt which is very much consumer land. since codex is intended for work & enterprise (which requires sales) it may not be an issue since it doesn’t have to grow organically as much. i’m curious if they ever combine the two in some way. obviously codex code is likely the basis for the new chatgpt app(s) / 90% shared code or whatever. anthropic does have a much cleaner story here although not as large of a consumer base.

It’s legit nice how well Codex and Claude Code get along with each other. They know each other’s strengths and weaknesses and buddy up. I have them review each other’s specs, design docs and code. They do it without ego or jealousy or insecurity - in a way that I haven’t seen with human engineers.

@signulll This is a banger post and the reason to be on X!! I have thoughts here but first wanted to just tip my hat at you sharing this.

the future interface is probably three layers: 1. ambient intent capture voice, location, calendar, screen context, messages, habits, biometrics, etc. the system understands what you’re trying to do before you explicitly “open” anything or augments your intent deeply. 2. agentic execution the actual work happens through agents operating software, apis, browsers, documents, email, calendars, workflows, payments, support systems, whatever. most “computer use” becomes machine to machine clerical labor. 3. ephemeral verification ux humans still need to inspect, compare, approve, edit, reject, or enjoy things. that’s where gui survives but as disposable, task specific surfaces generated for the moment.

@signulll Gracias, senor!

@amarg777 will dm you !!!

the iphone era app was a noun which loosely was a bounded surface you opened, performed an intent in, & closed. notes, reminders, clock, calendar, etc… these are all just typed buffers with light affordances around a primitive (text, time, date). the app was the price of admission cuz the os couldn’t infer intent, hold context, or act on your behalf. the icon grid was a workaround for a dumb constraint. once you have an ambient layer that knows you incl emails, calendar, location, health, finances, conversations.. the noun collapses into a verb. e.g. you don’t “open notes,” you mutter a thought & it gets routed, tagged, surfaced when relevant. you don’t “set a reminder,” the system notices the commitment in your text thread & pre stages it. time becomes contextual too (“you should leave in 12”). calendar becomes a derived view, & not an input first static surface. all of these experiences will have to be reinvented from first principles with an intelligence layer at the core not as something bolted on.

@signulll Give me access, yo! Signed up when you announced it.

@amarg777 skyeapp.ai

@Mr_Husky1 Our son called it Fran Sansiko. SF will always be Fran Sansiko for us.

Admit it: Your kid mispronounced a word three years ago and now the whole family says it that way. What's the word in your house? (We still say 'pasketti' and I’m not stopping anytime soon.)

@businessbarista Claude

I want to start a community dedicated to Claude Code. It’s become the gateway drug to coding and experiencing the power of AI for tons of people. This will be a space for people to share killer use cases, agentic workflows, proven prompts, and connect with other CC obsessives. Comment “Claude” if you want to join.

@elonmusk This is JK Rowling level name. Up there with Knockturn Alley. Do it!

Moonbase name: Auda City Or is too luney? 🤔