Tag: Jason Calacanis

A Dreamer’s Idea: Scaled-Down OpenClaw Agents on Smartphones Building a Private Enterprise Hivemind

Full Disclosure: Grok LLM wrote this for me at my behest. I could actually write something like this if I wanted to, but this is just for fun. Grin.

I’m just a regular person in a small Virginia town who tunes into the All-In Podcast and scrolls X a bit too much. No technical background, no code to show, no plans to build anything myself—just someone who finds certain ideas genuinely exciting and worth floating out there. I don’t have the expertise to make this real, but I think it’s a cool concept that could click for the right people once smartphone hardware and agent tech mature a little more.

Jason Calacanis’ recent energy around OpenClaw has been hard to miss—the accelerator push, the $25k checks for builders, the stories of people automating old jobs and turning them into leverage. It’s inspiring stuff. If this post ever reaches you, no pitch or ask here—just a simple “what if” sparked by your enthusiasm for open-source agents that actually do things, combined with Chamath’s ongoing point about enterprises hesitating to send proprietary data to the cloud.

The core hesitation is straightforward: cloud AI is powerful, but it means uploading sensitive info—customer data, internal strategies, trade secrets—to someone else’s servers. Latency adds up, costs stack, and control slips away. Sovereign AI, keeping data and intelligence inside the organization’s walls, feels more urgent every day.

What if we took the spirit of OpenClaw—the open-source, autonomous agent that runs locally, handles real tasks via messaging apps, and grows through community skills—and imagined a scaled-down, lightweight version running natively on employees’ smartphones?

Call it a conceptual “MindOS” layer (just a placeholder name). These pocket-sized agents would live on iPhones and Androids, using the neural processing units already built in:

  • Most of the time (~90%), the agent focuses locally: quick, private tasks like summarizing notes from a sales call, analyzing CRM patterns offline, drafting responses, or spotting anomalies in personal workflow data. No data leaves the device unless explicitly shared.
  • A small slice (~10%) connects to a secure company mesh over VPN—peer-to-peer style, sharing only anonymized model updates or aggregated insights (think federated learning basics). Raw proprietary data stays put; the hivemind grows collective smarts without exposure.

Cloud vs. Swarm in simple terms:

  • Cloud AI: Data goes out for processing. Great scale, but your secrets mingle in shared infrastructure.
  • Smartphone Swarm AI: Intelligence stays distributed across your workforce’s devices. Faster for real-time needs, cheaper (no constant API calls), resilient (no single point of failure), and private by design.

Practical angles for businesses:

  • A finance team gets better fraud detection as agents learn patterns across branches anonymously—no customer details ever shared.
  • Sales reps on the road pull instant, offline insights from deal history; the collective refines forecasting without cloud round-trips.
  • Healthcare or ops folks analyze notes or supply data locally; the hivemind quietly improves over time.

The longer-term appeal: This setup could let a company build its own evolving intelligence privately. Start with everyday automation, then watch the swarm compound knowledge from diverse, real-world device contexts. Unlike cloud models where breakthroughs get diluted or locked behind a provider, this hivemind stays yours—potentially scaling toward more capable, versatile agents down the line.

Smartphone hardware is heading that way: efficient quantized models, better battery management for background work, and OpenClaw-style frameworks already proving agents can run persistently on devices. Challenges like secure coordination and consistency are real, but solvable in an open ecosystem.

I’m not pretending to have the answers or the skills—just connecting dots from podcasts, your OpenClaw hype, and the sovereign AI conversation. If it sparks a “hmm, interesting angle” for someone building agents or thinking enterprise, that’d be neat. If not, back to listening and daydreaming.

OpenClaw #EdgeAI #SovereignAI #EnterpriseAI #AllInPodcast

J-Cal Is A Little Too Sanguine About The Fate Of Employees In The Age Of AI

by Shelt Garner
@sheltgarner

Jason Calacanis is one of the All-In podcast tech bros and generally he is the most even keeled of them all. But when it comes to the impact of AI on workers, he is way too sanguine.

He keeps hyping up AI and how it’s going to allow people laid off to ask for their old jobs back at a 20% premium. That is crazy talk. I think 2026 is going to be a tipping point year when it’s at least possible that the global economy finally really begins to feel the impact of AI on jobs.

To the point that the 2026 midterms — if they are free and fair, which is up to debate — could be a Blue Wave.

And, what’s more, it could be that UBI — Universal Basic Income — will be a real policy initiative that people will be bantering about in 2028.

I just can’t predict the future, so I don’t know for sure. But everything is pointing towards a significant contraction in the global labor force, especially in tech and especially in the USA.

Eat The Rich: Income Inequality Is Destroying American Democracy

by Shelt Garner
@sheltgarner

I’m listening to The Week In Startup and I’m flabbergasted in the mentality of Jason Calacanis and Molly Wood. I think some of my shock comes from something we don’t talk about a lot in America — class. They’re both really, really rich and so of course they would rather think up ways to screw over workers rather than pay people more.

A lot of America’s problems come from income inequality. If we would just tax billionaires out of existence and pay the average person a living wage then I think a lot of our problems would be dramatically mitigated.

But that’s just not what’s going to happened.

We’re going to use, on a macro basis, automation and AI to destroy a wide range of jobs that we assume will always be there. Off the top of my head, I find myself worried about the 3 million high paying transportation jobs that will vanish whenever Elon Musk hooks AGI to an EV semi.

One thing that was touched upon during This Week In Startups is that as long as what AI does is “just good enough” then it will be used. If you hook up AI to androids then virtually overnight the need for a Universal Basic Income is going to become clear.

A UBI solves a lot of problems and yet it would also cause a lot of problems. Most people are lazy and they just won’t do shit if they get a UBI. What’s worse, there are also people who are extremely ambitious and they would demand the right to have SOME job, just to be able to have somewhere to go during the day and have some sense of self-fulfillment. So, it seems as though some sort of “value added” stipend would be included in which if you can do some sort of job that AGI can’t do very well then you will extra momney.

Or something.

The point is — I’m really worried about America’s instability going forward because of income inequality. There seems as though there might be a real risk of the “Petite Singularity” causing a huge pushback involving the rise of neo-Luddites.

These neo-Luddites will demand huge carve outs whereby only humans can do some jobs. Only time will tell.

A Hot Take On The ‘All-In’ Podcast

by Shelt Garner
@sheltgarner

I’m growing fond of listening to the tech bro podcast “All-In.” I generally don’t agree with a lot of what they say and their built-in bias to one Elon Musk. And, yet, given that I am the red headed step child of my family politically, it is useful to build up mental antibodies to their version of the “mind virus” I keep hearing about.

Sometimes what these well heeled tech bros DON’T say is just as interesting as what they do say. Take, for instance what happened in the most recent episode of the podcast. Jason Calacanis obviously went to the Dave Chappelle show where Elon Musk was booed…but he neither mentioned that he went with Musk, nor mentioned that Musk was booed.

But, as I was saying, I like a good debate and I also believe that any thinking person worth their salt should bounce up and down on what they believe, challenge what they believe, on a regular basis so they have a better sense of what it is, exactly, that they believe.

I have to admit that I even, on occasion, find myself agreeing — to an extremely limited extent — with the podcast’s resident MAGA-friendly person David Sacks. But I think that says more about how I come from a Traditionalist background and am more anti-MAGA than I am any sort of liberal. And usually whenever I “agree” with Sacks it’s on the edges where simple common sense comes into play.

Though, as an aside, I have to chuckle at how the tech bros of All-In constantly and consistently conflate the hard power of the government they want so bad with the soft power of the “woke cancel culture mob”…that hurts their fee-fees? People like Sacks, specifically, really want to burn everything to the ground because they find “libtards” hypocritical and annoying.

So what, is all I gotta say.

Yes, Biden has turned out to be far more of a progressive than a lot of people expected, but he’s good at retail politics and sometimes you gotta dance with the one that brung ya. And the MAGA hatred of the “woke cancel culture mob” all goes back to how its a movement based on hate. They’re fucking Nazis dressed up in modern garb with a thin veneer of civility and indignation shellacked on to them.

A Nazi is still a Nazi. And I fucking hate Nazis, be they MAGA or old school in nature. Fuck MAGA Nazis.

Anyway, if you’re a self-aware, thinking person, I encourage you to challenge what you believe, to get out of your silo and listen to podcasts that don’t follow your personal media narrative or reside within your social media echo chamber. You’ll keep your mind healthy.