manveerc

Totally agreed. I sometimes wonder if they are making the model "lazy" with each iteration, it keeps getting better at avoiding work.

Congratulations on the milestone! Arcade is the best product for solving reliability and security in the action layer. It’s a no-brainer, given the team.

And time to value is another dimension. In your case finding the right person, scheduling the project, you implementing and delivering at best is a week if not more. With AI they get it in seconds. I may have fudged the numbers but the scale relative gains likely will be same.

WhatsApp had real network effects built in, and network was the moat. Don’t think Cursor has any real moat.

Well i built an equivalent of OpenClaw using Claude Code and hooking it up with WhatsApp. For mew I'm currently using it for following things

1. Morning brief + meeting preps 2. Managing client work and action items (tracking status, deliverables, etc) 3. Executing our AI workflows on my laptop. We have built several AI workflows for our agency and this setup gives the ability to seamlessly execute and control them through both mobile and desktop

Next on my to-do list is to build additional workfows for me and my wife around family logistics (travel, childcare, etc)

Are you asking if Agents should use API?

In my opinion sites that want agent access should expose server-side MCP, server owns the tools, no browser middleman. Already works today.

Sites that don’t want it will keep blocking. WebMCP doesn’t change that.

Your point about selenium is absolutely right. WebMCP is an unnecessary standard. Same developer effort as server-side MCP but routed through the browser, creating a copy that drifts from the actual UI. For the long tail that won’t build any agent interface, the browser should just get smarter at reading what’s already there.

Wrote about it here: https://open.substack.com/pub/manveerc/p/webmcp-false-econom...

Oh yeah I agree it’s bad. I just meant, company has no morals. It is very data hungry and doesn’t care about people’s privacy.

And with respect to past tense, I don’t know if they still do when they were caught red handed about some of these things. Unless there is a court order I am sure they still do, but I have no proof point.

Not something new. They were recording audio on Facebook app and messenger for the longest time without a people using the microphone. They were tracking people using network data. The list is pretty long.

Thats a good question. I would recommend MCP for the bulk of 'chatty' soft data to keep the database clean. However, you should selectively ingest 'high value' data into ClickHouse for vector search.

For e.g. you wouldn't ingest every 'good morning' message. But once an incident is resolved, you could ETL specific threads (filtering out noise) and the resulting RCA into ClickHouse as a vectorized document. That way, the copilot can recall the solution 6 months later without depending on Slack.

The interesting part is that only one of them is software only. I get it is economists but I guess this is also telling that while AI is really talk of town in silicon valley, long term it is one, minor if I may, part of the future.

Yes! And the layoffs

https://www.tryzenith.ai

We are a service to help brands navigate the new world of AI agents. Currently focused on helping them increase visibility in AI search but we plan to go beyond that.

As someone building a service around AEO (tryzenith.ai), I am curious how are you connecting which prompt drove traffic? No AI engine provides that

Thanks, will keep this in mind for future. Unable to change now.

Unlike most prompt injections, the researchers said Shadow Leak executed on OpenAI’s cloud infrastructure and leaked data directly from there. This makes it invisible to standard cyber defenses, they wrote.

Yeah the head line was clickbaity

https://archive.ph/QD3s5

If only this becomes an accepted industry wide practice as against 9-9-6

Anecdotally what I have seen is that these (tech) companies are not hiring much in US and if they are hiring, its in offshore locations. So yes 2025 seems better in terms of fewer layoffs but it doesn't really help the overall situation in the US.

I was trying to plan for some data analysis. Tried it few times with GPT5 Pro, it failed almost every time. It only succeeded once. While I didn’t had any issue with Thinking model.

Wonder what’s the cause of decline in views. One plausible reaction I had was that views might be down because of people using AI search (ChatGPT, etc) which unlike Google don’t show videos prominently. But since likes haven’t gone down that doesn’t seem likely.

For me 9/10 requests with GPT-5 Pro failed for some weird reason. This never happened with previous models. I ended up downgrading my subscription, I realized I wasn’t using it enough. And for me thinking mode has been good enough.

Referring to software and hardware talent, which are the biggest target for this

While I understand the rationale, is there enough talent in US to make up for the theoretical additional demand this can generate?

That’s more elegantly put than I ever can.

Btw I am not disagreeing with the utility of LLMs, my point is it can never be 100% accurate with current architecture (unless you blow up the size).

Maybe I am oversimplifying it, but isn’t the reason that they are lossy map of worlds knowledge and this map will never be fully accurate unless it is the same size as the knowledge base.

The ability to learn patterns and generalize from them adds to this problem, because people then start using it for usecases it will never be able to solve 100% accurately (because of the lossy map nature).

Hello hello Nikrad, it’s a well written blog. Love the authenticity:)

When I read the title, I thought you were referring to https://parallel.ai, which also is a game changer in my opinion :)

PS: I have no affiliation with Parallel the company

Your comment hits on a broader tension I see a lot, not just here but in business strategy in general. It's the divide between compelling, experience-based narratives and empirical evidence. I think both are essential.

The author has presented a fantastic and intuitive narrative with the "BUYER-PULL" model. Your analogy is spot-on: you can't sell 16 meals a day to someone who only needs three. The qualitative insight is powerful.

My request for data comes from the next step. How do we know this narrative is not just a "just-so story"? How much does this effect matter on the margin? In the complex world of B2B sales, where needs aren't always as clear as hunger, can "push" tactics sometimes be effective at helping a buyer crystallize a latent need?

Asking for metrics like close rates isn't meant to demand an impossible standard of scientific proof. Instead, it's an attempt to test the boundaries of this framework and understand its real-world impact. Great insights often come from quantifying the effects of a powerful story.