What’s the *real* way to deploy on Google Cloud?


I was talking to a customer recently and they asked about deployments. They wanted to do things the real way. Here’s a snippet…

I’m helping out a company called Blue Marble and they are getting ready to deploy a new POS system. The app has been built using a Node.js back-end and Google Cloud Datastore for storage. The current dev build is hosted on AWS and connects to Google for the data bits.

Join 38,000 others and follow Sean Hull on twitter @hullsean.

For prod launch, they are interested in migrating to the “real” way of deployment on Google for everything.

They are pressed on time and looking for someone who can jump in quickly. Are you available? Do you have Google Cloud expertise?

Here’s what I said.

Cultural hurdles

Yep, I’ve have used Bigquery & GCE.

What are they looking for specifically? Full deployment automation? Multiple deploys per day?

I’ve found that sometimes the biggest hurdle to fully automated deploys can be cultural issues.

In other words yes you can automate your deployment so it is push button, get all the artifacts & moving parts automated. Then deploy without much intervention. But to go from that to the team having *faith* in the system, that is a challenge.

Also: Why would I help a customer that’s not paying?

Unit testing

Once the process has been streamlined, a lot often still needs to happen around unit & smoke tests.

If the team isn’t already in the habit of building tests for each bit of code, this may take some time. Also building tests can be an art in itself. What are the edge cases? What values are out of bounds?

Consider for example odd vulnerabilities that show up when hackers type SQL code into fields that devs were expecting. Sanity checking anyone?

Read: Is AWS too complex for small dev teams? The growing demand for Cloud SRE

Integraton testing

What makes this all even more complicated is integration testing. Today many application use various third party APIs, service-based authentication, and even web-based databases like Firebase. So these things can complicate testing.

Related: How to build an operational datastore on Amazon Redshift with S3

Getting there

Although your project, startup or business may be pressed for time, that may not change the realities of development. Your team has to become culturally ready to be completely agile. Many teams choose a middle ground of automating much of the deployment process, but still having a person in the loop just in case.

Same with testing. Sure automating can make you more agile & more efficient. But you’ll never automate out creative thinking, problem solving & ownership of the product.

Related: Why did Flatiron School fail?

Get more. Grab our exclusive monthly Scalable Startups. We share tips and special content. Our latest Why I don’t work with recruiters

What does the failure of Flatiron School mean for coding bootcamps?


If you missed the news, New York’s AG announced a settlement with Flatiron School over operating without a license and false advertising.

The news of the coding bootcamp failure splashed all over hacker news a couple of days ago.

Join 37,000 others and follow Sean Hull on twitter @hullsean.

With the explosion of coding bootcamps in recent years, it speaks volumes, as demand for coding & software skills continues to outpace supply. What’s more the starting salaries aren’t bad either.

But how will this affect coding bootcamps going forward?

Would you like a helping of beaurocracy?

Part of the ruling was regarding licensed teachers…

“In order to obtain a SED license, a non-degree granting career school must meet a number of criteria, including using an approved curriculum and employing a licensed director and teachers.”

One thing that sets coding bootcamps apart is that they train their own teachers. And they also use their own curricula. And while protecting consumers is certainly a worthwhile goal, the ruling means bootcamps will have to navigate government bureaucracy for approval. Some have pointed out that the process can be slow & full of red tape. Which is sort of counter to the whole agile startup private industry philosophy. We’ll see!

Also: Is Amazon too big to fail?

$75k after 3 months?

One of the claims their marketing made was that many students were making $75k after a few months of study. The ruling underscored this as particularly misleading. more here

As anyone who has studied computer science knows, there’s a lot of foundational concepts in logic, mathematics & problem solving, which you don’t develop overnight. Hopefully this ruling with hammer home the idea, that it takes a little bit more time folks!

Read: Is AWS too complex for small dev teams? The growing demand for Cloud SRE

Does it please the crown?

One of the comments on hacker news asks “Does it please the crown?”. By slapping these guys on the wrist, the barrier to entry will be higher. Going forward, they will have to pass more hurdles & government beaurocracy.

One of the things that sets coding schools apart is that they can train their own teachers & build their own syllabus. We’ll see if these new hurdles slow things down or not.

Related: How to build an operational datastore on Amazon Redshift with S3

Billion dollar Google bet

Almost as if to provide a counterpoint, timely knews comes out of the google camp. They have commited to $1 billion in grants to train workers.

It seems demand for coding skills will be strong for the foreseeable future.

Related: How to build an operational datastore on Amazon Redshift with S3

Get more. Grab our exclusive monthly Scalable Startups. We share tips and special content. Our latest Why I don’t work with recruiters

Why would I help a customer that’s not paying? The reason might surprise you


I just received an email. It was from a woman building a website, and wanted help with AWS. She wondered if I might be able to provide any assistance.

Join 37,000 others and follow Sean Hull on twitter @hullsean.

Having a popular publicly facing blog, I get a lot of leads that seem to come out of thin air. This is the good problem of publicity. 🙂

I followed up with her and asked what she was building. “Nothing” she explained, I just want to learn about AWS. I was a little confused at first, but as we talked further, it seemed she was just beginning to branch out onto the wild world of the internet, and didn’t know where to start.

I explained that to build an e-commerce site, she could use a service like Shopify, and would likely not need to use AWS directly, and certainly wouldn’t have to learn it. That might take five to ten years learning computing first!

I realized I was telling her she didn’t need the services of someone like me, and further giving her half of a solution. Though I couldn’t help her build a product, the information could surely help her sell it.

Then I thought to myself, why would I do that? Why give away your time & advice for free?

1. Find time to followup

LESSON: a quick call is always worthwhile networking

Yep it’s true, I’ve learned over the years it’s always worth your time for a quick call. I even talk to recruiters on occasion though I don’t work with them.

You’d be surprised how often you learn from someone, especially when they don’t work in your domain. You learn from the way they frame questions, how others might view or search for you. You learn how better to explain & sell your services to future customers too.

Also: When clients don’t pay

2. Be helpful

LESSON: Provide some real help or value

In a call like this one, it costs me very little to “drop some knowledge” as the cool kids like to say. 🙂 Sure my time is worth something, and yes I’m giving something away for free. But in this case it was someone who currently doesn’t have the budget for my services so isn’t my target audience anyway.

Read: When you have to take the fall

3. Pay forward

LESSON: Always be networking

Be patient. As Keith Ferrazzi likes to say “Never Eat Alone”! I’ve taken hundreds of calls like this one over the years, and some later get funded & call me back. They’re eager to put me to work, already sold on my integrity & personality.

What’s more she may run in different circles than I do, bump into a colleague or recommend me at some point. If your openness really stands out, it’ll leave a memorable impression long into the future.

In a place like New York where we’re often singularly focused on profit & personal gain, it’s easy to stand out by a small act of kindness.

Related: A look at the serverless hype cycle

Get more. Grab our exclusive monthly Scalable Startups. We share tips and special content. Our latest Why I don’t work with recruiters

10 things you definitely didn’t know about bitcoin


The news is ablaze with stories of bitcoin. Everyday I see something new, whether it’s wild volatility, to some fraudulent transaction or other.

Join 35,000 others and follow Sean Hull on twitter @hullsean.

I’ve been digging into them myself of late, and learned a few surprising facts about them. Read on!

1. It’s disrupting banking itself

That’s right, it’s bitcoin is not just an experiment anymore. Cyptocurrencies (of which Bitcoin is just one example), are ruffling the feathers of the IMF. Managing Director Christine Lagarde predicts the end of banking. Now that’s pretty big.

Meanwhile Chris Skinner outlines the major use cases for blockchains. These are the scaffolding of digital currencies, the financial plumbing if you will. Among those uses include smart contracts & assets, digital identity, clearing & settlement & of course payments.

Also: Is Amazon too big to fail?

2. There are many digital currencies

Although Bitcoin was the first, and the one you hear about in the news a lot, there are many more. Etherium is one. Take a look at this colorful map of all the current cryptocurrencies out there. There are hundreds!

Related: Was Fred Wilson wrong about Apple?

3. Banks are already using it

UBS-led 6 bank clearing house initiative

Read: Why the android ecosystem is broken

4. You can mine them & spend them

Coins can be mined like gold in the real world, and of course spent using a digital wallet.

A. Mining: This requires either a physical device or a contract with a provider such as Genesis Mining. The more coins in the world that are mined, the harder it becomes to discover more. My early results show that, if the price remains stable, my contract will become profitable in about 18 months. In other words, I will have ‘made my money’ back on the original upfront price to get started. From then on it’s profit, but not stellar… about 20% per year. If the value of the coin itself goes up, it becomes more profitable, if the value of coin goes down, well then mining pauses, or ceases entirely.

B. Wallets: You get a website and a phone app, that allows you to pay/receive person-to-person. One thing I noticed is that there is a LOT of security, because they must be a target of a lot of hacking. My online bitcoin wallet, CoinBase provides not only 2-factor authentication (text to phone) but after your account reaches $1000 USD value you unlock a ‘vault’ in which you can store coins. You don’t earn any interest, instead you get extra security: you must verify your withdrawl/transfer request from TWO different email addresses.

Read: 30 questions to ask a serverless fanboy

5. You can build your own

With Build a Coin project you can build your own currency. This isn’t building an app for Etherium or an API to some existing currency. This project helps you build the code to support your own virtual digital fiat currency!

Read: Professional services and the art of resistance

We’ll be back soon with 5 more. Don’t worry!

Get more. Grab our exclusive monthly Scalable Startups. We share tips and special content. Our latest Why I don’t work with recruiters