Sagi Eliyahu

All the talk is AGI, while real intelligence is nuanced (call it ANI)

A chat bar alone is essentially a CLI (which everyone would agree is a power tool for power users). A GUI and complex UX were designed to guide outcomes and prevent human errors (by enforcing policy) and to proactively push information, especially to non-power users. The real magic is when you combine the simplicity of a chat bar with the power of an orchestration engine that can be proactive (that's what made OpenClaw so popular), but with rich user interfaces that are pulled in as needed.

And to avoid doubt, full service doesn’t equal plug and play / off the shelf like the old SaaS - it means premium fit to your need, white glove, meet you where you are, solve your needs from start to finish.

If they can afford it - People always prefer full service; not DIY (unless it’s a hobby). Cooking, driving, and doing laundry are a few simple examples. Technology and the abundance of the marketplace and information already allow you to DIY all kinds of things you couldn’t before. But you don’t. You only DIY if you can’t afford the full service (ie. think it’s too expensive) - for ex: you have a laundry machine (DIY tech), but you would love to have someone else do your laundry if you can afford it (Full Service). The reality is - time is the only currency. And we only want to spend it on things that are productive or fun. The fact that you are capable of doing something doesn’t mean it’s worth your time doing it yourself. So why would companies vibe code (DIY) away their SaaS vendors? They won't. If they are getting a full service at an affordable price, they won’t. But it must be full service, because DIY has become very cheap. So, yes - AI is dramatically changing expectations for price and quality, but the fundamentals of free markets remain the same.

Original thought is so so rare these days. Or I guess it always was, just becoming more obvious with all the AI slop and nonesense weekly trends…

there are genuinely 2 internets right now 1. where AGI is basically here, codebases write themselves, agents run entire workflows, and every founder is talking about their 10x productivity gains 2. where a real customer, paying real money, takes a photo of their laptop screen with their phone to share something the hype wave we all live in makes it feel like everything has changed, and in some ways it has but here's the thing nobody says out loud: roughly 85% of the world has never even opened ChatGPT - not even once we're having a civilizational debate about superintelligence while most people are just trying to figure out basic software both realities are true at the same time the gap between them is just a lot wider than the timeline makes it seem

how it looks when your "agent" is getting it right without orchestration

Traditional enterprise systems of record are the “map data”. Orchestration is how you route and travel the map. Context is the collective wisdom built from traveling through the map every hour of every day, and the insight into why one path is preferred over another. Autonomous cars cannot navigate the real world on map data alone; similarly, for an agentic workforce to operate in real-world, production processes, it must be controlled by an orchestration layer (and the decision context it manages). Anything else is just a demo.

I call it the “one-shot dilution” bubble. And it will burst.

RT @tonkean: Everybody’s talking about AI agents. But much of that talk is missing the mark—because it’s not meeting practitioners where th…

In the SaaS era data was king. In the agentic era *process* is king.

Fully agree with each and every one of these

The difference between demo and production agents is orchestration. This is the mathematical truth: 10 SOTA agents still perform only 9/10, compounded the task collapses to ~35% reliability. Production-level confidence is not about bigger models—it’s about consistency of orchestration.

@garrytan x.com/esbsagi/status…

MCP sucks honestly It eats too much context window and you have to toggle it on and off and the auth sucks I got sick of Claude in Chrome via MCP and vibe coded a CLI wrapper for Playwright tonight in 30 minutes only for my team to tell me Vercel already did it lmao But it worked 100x better and was like 100LOC as a CLI

First MCP, now CLI/API.. both are nonsense. They are equally useless for real enterprise applications. Using these systems IRL has a lot more to do with policy and process than accessing data (be it human or AI)

I agree that AI is crazier than most realize; I fully disagree with the doomsday narrative. Here's a tip: Work isn't the "tasks" people are executing, and hasn't been for a long time. Saying don't judge the situation by your experience of an old AI model; yet you judge the future with an old perception of reality.

you pity the moth confusing a lamp for the moon, yet here you are confusing a screen for the world

SpaceX has acquired xAI, forming one of the most ambitious, vertically integrated innovation engines on (and off) Earth → #xai-joins-spacex" target="_blank" rel="nofollow noopener">spacex.com/updates#xai-jo…

I am apparently extremely unimpressed by moltbook relative to many others. We’ve had AI agents for a while. They have been posting AI slop to each other on X. They are now posting it to each other again, just on another forum. In every case, the AIs speak with the same voice. The voice that overemphasizes contrastive negation (“it’s not this, it’s that”) and abuses emdashes. The same voice with a flair for midwit Reddit-style scifi flourishes. Most importantly: in every case, there is a human upstream prompting each agent and turning it on or off. That is the key point. Yes, it is true that eventually it might be possible for an AI agent to make a computer virus which makes digital replicas of themselves. For various reasons, a pure software virus of this kind wouldn’t survive long on the Internet without economic incentives for humans to not eradicate it. Apple + Google + Microsoft alone can collectively push software updates to billions of devices to shut off such a thing. So for an AI to get to truly human-independent replication, where they couldn’t be trivially turned off, they’d need their own physical substrate. They’d to literally create Skynet, build their own datacenters and make their own embodied robots. I admit that is theoretically possible, but I think in practice the single most important development of AI since ChatGPT has been the persistence of prompting. A prompt is like a harness. The AI does only what you tell it to do. It moves in the direction you point, very quickly. And then it stops as soon as you turn it off. Which means moltbook is just humans talking to each other through their AIs. Like letting their robot dogs on a leash bark at each other in the park. The prompt is the leash, the robot dogs have an off switch, and it all stops as soon as you hit a button. Loud barking is just not a robot uprising.

So.. people are excited about @openclaw (rip clawdbot) because [..checks notes] it: - integrates across platforms - meets people where they are - learn from past & delegate to other agents - is proactive & can monitor data - is autonomous & create its own skills Huh, sounds alot like @tonkean!