Thread by @PaulDJohnston

Some of the people I respect the most as CTOs are not even on Twitter.

Follower numbers, retweets and likes are a terrible way of working out whether something is good or bad.

As is falling for the messaging of large marketing budgets.

Trust isn't gained by GitHub stars or which vendor you work(ed) for our where you went to college.

Trust is gained by whether you've actually done a thing and can bring real world experience to the party.

e.g. I've been a CTO of a serverless startup

That gives me credibility

Credibility means that you can trust that I vaguely know what I'm talking about

The fact that I have worked for Amazon as a Senior Developer Advocate for Serverless... for me holds weight only because of the link between customers and product.

I heard the problems and solutions

So not only do I have my own experience but I have inside understanding of multiple companies and how they use serverless techniques to achieve their goals.

It means I've seen most of the mistakes as well.

So I can write a Serverless Best Practices from a position of some insight

And I have...

bit.ly/serverlessbest…

But I don't have the follower numbers because I don't speak at *every* conference

I've only worked for one vendor and didn't speak at that many for them

And I've never worked for a tooling company or similar so never needed to speak at conferences then either

I only started to speak about serverless because it was so new and because I was looking to hire people

As the CTO of a new startup with no access to talent where do you look?

It sort of worked.

It made me friends (hi @thomasj and @IamStan and @ben11kehoe) and others

And I wish more people wanted me to consult on serverless stuff

But I'm not seeing the demand

I'm seeing the demand for AWS consulting - EC2 and the like

I'm seeing the demand for Kubernetes

I'm not seeing the demand for serverless

Interestingly, most people see it as easy or trivial, or an extension of microservices...

Most people think you don't need a consultant...

Most people can get away with it for a while...

But I've seen most of the things that go wrong and they are often the same

Usually it's RDBMS based or it's using Java and some massive framework with really awful cold starts or it's thinking that you have to use a VPC or not understanding how to construct your data layer or constructing everything as microservices and having lambdas calling lambdas...

None of these things in and of themselves are wrong in certain circumstances, but they shouldn't be your initial starting points...

And that's the thing.

I know this stuff.

But nobody wants to pay for this knowledge up front.

I've been a CTO

I've worked at AWS in this space

I know this stuff

And yet most people are happy to repeat the mistakes of everyone else without asking for even a basic level of help

Half the time they haven't even read my blogs or any of the serverless community's blogs

I just don't get it

When I started, nobody was doing this

I looked and had to figure stuff out for myself

Because there were no blogs or companies to follow

But there are now

If you are thinking of going serverless... And you don't quite know where to start...

DM me

I'm pretty sure I can save you time

Mostly in the fact that you'll start right and won't have to refactor anywhere near as early as you will if you don't

Happy to help

Why did I start out talking about GitHub stars and follower numbers?

Because we all like a good retweet and to point at numbers and metrics

There are many projects with lots of stars in GitHub that I would never use

Lots of followers on Twitter doesn't mean they're better

And you don't have to agree with me

Please make up your own minds

But I'm really bored of this whole serverless game where vendors claim that something is serverless

The word has become meaningless

I don't do Kubernetes - it's not serverless to me

I do Lambda and AWS