I have always considered SEO the most fundamental skill for designing websites.
Broadening a bit into seo + aeo + websites + content + cro + web analytics is a good option. Especially as roles compress.
Hell we are not so far away from them + their claude code being an interesting option for owning web dev.
Why SEOs don’t universally own IA is one of the most bizarre patterns in the industry IMO.
Right after we officially decided microservices were a bad pattern for apps we turned around and made agent microservices.
The architecture is the org chart.
This by Anthropic seems both right & a condemnation of many “agentic solutions” & harnesses being built at companies. Rather than a sprawl of specialized subagents built by different teams, you often get better & cheaper results using one agent to do it all https://t.co/0dkBHZzUiU
The AEO has nothing to do with SEO ppl bother me the same amount as the AEO is just SEO ppl.
The AEO is just SEO ppl are the more correct of the two but I don’t care.
Both are just talking their book and biases.
Neither can be a reliable steward of a brand.
Just get one claude session to manage all your other sessions. Including the ability to create and close existing ones.
Use crons that regularly (every 15-30 min) ping that orchestrator and nudge any idle sessions.
Ask the agent how to get rid of all the useless permission requests, use 1pass service accounts, etc.
It is quite possible now to give agents a week of work and (mostly) walk away.
Long-term there’s no difference between vibe coding and coding.
If you don’t think vibe coding is a worthwhile endeavor what you are saying is that you don’t think building software is valuable.
And before you mock someone new to software for vibe coding a $99 SaaS, first consider that doing small well-defined projects, creating software people might actually use, is as good a way to learn as you will find.
Building calendly replacements is probably not what you want your growth team doing full-time. But as a 20% project where the intention is as much development as it is the explicit output?
Great. Go for it. You will learn a lot.
And your intuition for what to do and build (and what not to) will improve the more you do it.
Updating the parts of your ui that aren’t reachable over mcp (certain stuff in settings in particular) is an increasingly complicated choice.
Maybe you made it slightly better. But now not only can an agent not reach it, it has wrong knowledge about where it is or how to do the thing.
The net of which is making agents an unreliable user of your product.
The ux improvements better be worth it.
Execution mostly.
Most of how big companies scale is by putting more products/services into a small number of GTM motions that fit together.
The higher level strategy work is critically important to get right but the challenge there is not a high volume of work. It is the high judgment work of knowing what to do. And does not necessarily improve as you add more people to it.
So real domain expertise is still required. But building systems is how people will turn that expertise and judgement into professional leverage.
On some level this has always been the case. But traditionally the way to develop that leverage was through things like leadership, management and communication. Because the way to scale was by building big, effective, cohesive teams.
But now some large percentage of how you scale will look more like an engineering problem than a traditional mgmt problem.
And so much of being a leader will not just be about building teams that can do the work but systems that can do the work.
The idea that ai use will be something like “undifferentiated tool use” is so absurd.
There is a mountain of difference already and that gap is widening.
If you can’t see that it is a failure of imagination.
You are failing to imagine how you become a team of 100 or 1,000.
Could I manage all digital marketing for a fortune 500 by myself?
So much of what we do is develop best practices and sops, teach them to people and monitor the results.
Why can’t I just program all of that instead? I think a million lines of code will cost about $10k in 2027.
With zero meetings I could also do meaningful review/feedback on 100- 200 decisions a week.
Are digital teams making 5,000 genuinely important decisions a year? They are not.
I’m also not saying I actually want to do this by myself in my ideal Future of Work scenario. I like working with people. I think this would be lonely and boring.
But I think it’s a useful provocation for imagining how things will be different.
And for why being able to build ai systems will be more important than all the other things.
Because it is where nearly all the leverage is.
@gaganghotra_ I don’t think these metrics are a surprise to anyone. We have ai referral traffic.
Most ppl understand you have to measure as a brand channel.
There’s been more cynicism about the value of vibe coding on the timeline but that is extremely short-term thinking.
It is very clear that we will have GTM systems/factories.
And that the highest leverage thing you will be able to do in marketing is build and run that system.
But creating that requires a much deeper understanding of both software and infrastructure than most marketers possess.
The best way to learn how to build systems is to go build some systems.
Each specific thing you build may or may not have value. But that is barely the point right now.
If you are measuring personal ROI of each app, workflow and agent you are mostly missing the point.
The real question is how relevant and valuable your skills are for where the discipline is going.
And whether you think every role needs to be completely ai-pilled is beyond the point (even though the answer is that they do), the point is that a very high % of the work will ultimately be done by agents.
Do you know how to build a system, and a function, where agents are doing much of the work?
If you don’t build a lot of things yourself, how could you?
And if you can’t do this, how can you lead a marketing function?
So ya you should try to build things that are ambitious and valuable but don’t forget that part of the journey is just learning how to build.
And that’s long-term not just valuable but maybe existential.
So ya go forth and vibe code that $99 SaaS.
The inbound era was a truly magnificent and perhaps still under appreciated era of marketing.
Enormous channels
Very high quality data
Reaching ppl by creating value
Compounding, saas like economics
We may not get another one like that.
Introducing Claude Meetings: add Claude to your team’s video call and it builds docs, whiteboards, and other artifacts with you while you’re still talking.
Starting to roll out in beta on Team and Enterprise plans.
I somehow forgot to add, the greatest irony is that Jensen’s tactics are the complete opposite of the original mgmt bible of silicon valley authored by none other than Andy Grove, CEO of Intel.
So even in chips you have multiple historically great companies and leaders with polar opposite leadership styles and methods.
Today makes it pretty obvious why so many of Jensen’s unique mgmt tactics (extremely flat org, no 1:1s) are neither generalizable or even likely work for an NVIDIA he’s not leading.
There are a few important lessons here.
First, nothing matters more than being right about the most important thing. You can get a hundred small things wrong but if you get “we should make the best gpus” right it can cover for many other mistakes.
I would go further that many mgmt and culture practices, books, etc are mostly weak correlation with a great business. Netflix and Google and NVIDIA and even Spotify all had moments where everyone was studying their unique cultures but all these cultures were and are quite different.
None of them offer much in terms of practical or transferrable insight. Beyond maybe a couple bumper sticker ideas like the “keeper test.”
But taken together they do offer a type of meta insight.
Which is that great companies are often built around the unique operating styles of their founders. And there are many versions of what that can look like and be successful. One is not better than the other but what makes it work is that the company is shaped like a thing the founders can be successful with.
Probably everyone has had the feeling at some point of… if you let me work the way I work I and we will be successful. But if you make me be something I am not I will not succeed (at least to the same level).
Founder and CEO is that same thing but at the level where personal success and company success are synonyms. And if you are a founder trying to do new hard things you need an org that has a shape that feels organic to your ways of working, which then become company ways of working.
The key is having enough self-awareness to recognize the difference between generalizable insights and practices and unique ways you are building a team and rhythm that is designed to amplify your gifts and compensate for your weaknesses.
And the trap is building orgs that are so far shaped around the presence of a 1 of 1 leader that if anything happens to that person, the org will need to be massively transformed.
In which case the decisions you are making now (which might legitimately be helping you succeed today) are not practices to share, admire and learn from, but are rather a type of organizational technical debt.
It’s sometimes hard to know the difference between whats generally good practice or advice vs whats good for one founder in one company.
But it helps to appreciate how different even great companies can be.