Yass

A bit of AI. A bit of DeFi. A bit of philosophy.

[Yass @Yass_nl]

@AbdullahAd29959 I mean yeah it’s broad, but it remains one piece of CS among many others

[Yass @Yass_nl]

Why do some people still think that CS = software engineering ?

It may not even account for 15% of what computer science is.

[Yass @Yass_nl]

A thought on the potential future of Agentic AI https://t.co/ox8x2ZRRbN

[Yass @Yass_nl]

Top #1 on «Outage» benchmark as well haha

[Elon Musk @elonmusk]

@GavinSBaker @theallinpod Anthropic is #1 on “RevenueBench”, which is by far the most important benchmark!

[Yass @Yass_nl]

Multi-agents workflows are only as good as the agents ability to communicate with each other. Agree ?

[Yass @Yass_nl]

@bridgemindai At this point it’s just part of the routine. No problem

[Yass @Yass_nl]

Idea generation

[chanduu @chanduuu_cs]

AI can write.

AI can code.

AI can design.

What skill becomes MORE valuable because of AI?

[Yass @Yass_nl]

@NoahKingJr Instagram and TikTok is slop. LinkedIn is the hub of hypocrisy. X remains the closest one to reality according to me.

[Yass @Yass_nl]

Knowing that "there" is usually followed by "is", and that it switches to "are" when the word after ends with "s" ;

That does not make anyone ( or anything ; ) intelligent.

[Yass @Yass_nl]

@FeelipeFelten @LukeDadev Thanks for the info, if Reddit does not retract the ban I ll definitely reach out

[Yass @Yass_nl]

Got banned, can someone explain ?

I posted once on reddit where I laid out my negative view on the "Internet of Agents".

And after a few hours and few thousands of views ( basically nothing ), my account got banned.

Why ? Is it a mistake ?

( Text Below )

[Yass @Yass_nl]

@N_and_ni That’s a damn good question

[Yass @Yass_nl]

There is literally no reason to switch, at all.

[Jiya @jiya_3063]

I'm a Vs Code user.

Give me one reason to switch to Cursor.

[Yass @Yass_nl]

@HeyAliux To me Claude remains the best choice. Anthropic got me until the end of my life lol

[Yass @Yass_nl]

@anah_sahh The market I guess

[Yass @Yass_nl]

I d like to give my take on a narrative that's picked up a lot of momentum: the idea that we're heading toward an "Internet of Agents" — a shared layer where companies publish agents that other companies (or anyone) can discover and rent, presumably paying per token or per call. AGNTCY, ( Linux Foundation with Cisco, Dell, Oracle,...) is the most visible push in this direction — identity, discovery, messaging, observability, all built so agents from different vendors can find each other and interoperate across org boundaries.

I personally don't buy the endgame a lot of people are projecting onto this.

Just to be clear: interoperability standards for agents talking to each other, or to tools, have obvious value — MCP-style stuff solves a real problem. I'm not skeptical of "agents should be able to communicate using a shared protocol." That part makes sense to me.

What I don't buy is the sort of "marketplace of agents" framing — the idea that Company A will routinely rent and run Company B's autonomous agent inside its own workflows. A few reasons:

An agent isn't like compute or an API. Renting compute or calling an API is a well-understood trust boundary — you know exactly what you're getting and what it can touch. An agent is an autonomous decision-maker with its own goals, tool access, and failure modes. Letting a third party's agent operate inside your org is a much bigger trust ask than anything we've normalized so far.

A lot of the infrastructure here is solving a problem the premise itself creates. Identity, cryptographic verification, access control across organizational boundaries — these are exactly the features you need if you've already decided agents will be shared across companies. If that doesn't happen at scale, a meaningful chunk of this complexity doesn't need to exist. It's solving a self-inflicted problem rather than a pre-existing one.

Ownership, not rental, matches how companies already behave. Most enterprises are already cautious about running open-source models they don't fully control. Handing decision-making authority to another company's agent is a heavier lift than that. My bet is companies will keep building and owning their own agents in-house — even if slower and more expensive — because the liability and control stay contained. This mirrors how automation has generally gone: companies consume standardized data/APIs, but keep the actual decision logic in-house.

So my take: standardized communication between agents, yes. A functioning cross-company rental market for agents, not convinced. Feels more ideological ("open internet of everything") than something with real enterprise demand behind it.

I d like to know what you all think, maybe I miss something about this whole thing.

[Yass @Yass_nl]

@axhiii88547 IntelliJ ❤️

[Yass @Yass_nl]

Am I the only one who feels like Opus > Fable ???

[Arun @hiarun02]

@trevrjay Yeah 😆 being an Opus user right now is a weird experience. Fable just feels way more capable overall.

[Yass @Yass_nl]

App dev by far, it’s horrific

[Archna @archnarambles]

Be honest, which is more complex?

- Web Dev 💻
- App Dev 📱

[Yass @Yass_nl]

@Surendar__05 Cursor is a derivative of vs code, so cursor. But vs code for life anyways