HashiCorp Field CTO Weekly: The Quickest Honeymoon, Pipeline Solves All Woes, and a Call To Action
The Quickest Honeymoon
A few weeks ago I was discussing (like many people) the fast emergence and adoption of ChatGPT and related AI/ML systems. So I started digging in deeper for a much better understanding of the systems and field in general. I took the original class from Sebastian Thrun on how to program self-driving cars, which ultimately lead to the formation of Udacity, so I had a bit of knowledge of some AI programming. But that was 10 years ago and I needed some updates. I mentioned Andrej Karpathy in that previous newsletter and I was pleased to see he put out a very extensive crash course called Neural Networks: Zero to Hero. This course is quite good and I learned a lot about neural networks and even, to some extent, how to program them “the hard way”.
But then, the downsides started to show. ChatGPT was passing exams, ChatGPT was creating malware, other neural networks were creating deepfakes that were consistently fooling people. Our delight and amazement with our new found AI friends were quickly starting to dwindle and fears of the dystopian AIs crept back in. But I think my favorite article was ChatGPT and the Enshittening of Knowledge where the author Daragh O’Brien discovered that ChatGPT outright lied (in a convincing manner) about source materials it used to generate its output. Much like my claim that we need actual human input into the system for it to emulate us properly, Daragh said (warning explicit language):
When we turn the job of creating the first draft of the thing over to ChatGPT we risk removing the “figuring shit out” part of everyone’s career-path. And we get away with that while we still have people who have figured shit out through experience. But that expertise is time-bound. People move on. Unless we actively plan for knowledge management and skills development in an era of automagically generated first drafts, we’ll eventually get to the point where there will be nobody left to call out the A-Grade Bullshit generator when it is spouting shite.
I’m all for workflow accelerators, but if we want to progress without AI biased feedback loops, we need to continually produce human-made media.
(No AI/ML was used in the creation of this newsletter).
Pipeline Solves All Woes
I’ve spent most of my career in IT delivery of some form, most of it in systems administration/devops/platform engineering(before we even called it that), but also network engineering, support and professional services. But I’ve also done a few rounds of sales engineering positions over the years and I’ve come to appreciate both the energy and the jargon that comes along with being in sales. There are the inevitable jargons like TAM (total addressable market), ACV (Annual Contract Value), NDE (Net Dollar Expansion), and hopefully CSAT/NPS (Customer satisfaction/Net Promoter Score). But today, the one concept I want to focus on is Pipeline. Different folks will define pipeline slightly differently, but there are some common themes in that sales pipelines, much like software delivery pipelines, describes the different stages of the sales process from initial qualification of a lead (the folks who are potentially interested in buying our products) all the way to the sale of the product where the lead or prospect becomes an actual customer. And because not all prospects will become customers, sometimes folks associate the visual of a marketing or sales funnel to describe this. That is, at the top of the funnel are all potential customers (leads/prospects) and through the various stages of the sales and marketing processes we qualify less and less prospects until we reach the actual paying customers.
Because of this, it’s always highly encouraged to have as many prospects as possible at the top of the funnel at the beginning of the sales/marketing process. And that’s where the phrase “Pipeline solves all woes” comes from (there are other such catch phrases, but I’ve been watching Wednesday on Netflix and couldn’t help attaching a nod to the show). That is, if the top of your funnel is as large as possible, it’s more likely you’re going to convert the amount of customers necessary to make your quota.
It’s possible at this point you’re thinking “Why the heck is this guy talking about sales and marketing in a blog traditionally focused on technology?” To which I’d say “That’s an excellent question!” (which you’ve also probably heard from someone in sales).
During one of those phases of the sales pipeline shown above is “Lead Qualification” (sometimes called “Discovery”). It’s at this stage that we learn a lot about our prospects’ problems and how they impact their business and whether or not our products will actually help them solve both their technical and business objectives. There’s one common theme across the board when talking with prospects (and customers) and that is, that almost everyone has a talent issue.
There are many different factors involved in our talent shortage. One is related to the way we think about hiring into tech jobs in the first place. At the time I entered the work force in the mid-90’s, most entry level systems administrator jobs required a 4-year degree in computer science or electrical engineering or some other related engineering discipline. I had a B.S. in Asian history. However, because I had 8 months until I started pilot training, I had to do something, and the Air Force has something called “casual status” jobs where newly minted lieutenants get to do other jobs in the Air Force while waiting for Undergraduate Pilot Training to start. And this is where I got my first job as a Solaris systems admin working for the 10th Communication Squadron at the Air Force Academy.
Something interesting occurred to me much later after I spent time as a systems administrator as well as working in support (potentially more-so here), and that was my degree in history was actually very beneficial for those roles. Being a(n) historian involves reading and viewing lots of different kinds of source materials like newspaper articles, letters, pictures, artwork, music, clothing, etc and giving an analysis of what you thought was going on at the time. In its simplest form, it’s turning data into information. For anyone who has worked in support, this probably sounds familiar. Your customer sends you a bunch of data describing the state of their system, and you analyze it to find what the problem is. In many cases this process is iterative. It turns out my education as a(n) historian has helped me quite a bit in this career.
The reason I bring up my personal journey is because there are probably plenty of other examples that are similar out there. One of my favorite talks I’ve watched recently is “Transformed! A Musical DevOps Journey with Forrest Brazeal”. In between the hilarious musical reflections on things like re-orgs, Containers vs. Serverless and belonging, there’s a great intermission where he talks about the Cloud Resume Challenge. While I know a lot of folks don’t bother clicking on the links provided here, this is one I highly encourage folks to visit. For those of you that don’t want to, the tl;dw (too long; didn’t watch) is below.
TL;DW: The cloud resume challenge is enabling folks from completely different careers outside of IT, and helping them transition into entry-level cloud engineering positions. He begins with the story of a plumber who made this journey and applied skills he learned while doing plumbing to cloud engineering.
This is especially important when we look at the distribution of overall engineering graduates (~6%) here in the US according to College Graduation Statistics on the Education Data Initiative
While not all readers will be from the US and distributions will be different in your country, a lot of countries where our main paying customers are at have similar distributions for engineering degrees. But more importantly is that we’re not even capturing folks who might be coming from trades like plumbing, HVAC, electricians or even non-trades like hospitality workers.
Call to Action
No, I don’t want you to go setting buildings on fire or starting a fight with someone you don’t know and losing. What I would like you to do is listen to folks in your community, be they friends, family, associates or folks who help you on a regular basis. If they’re looking to change careers or are working in IT and want to level up, introduce them to the Cloud Resume Challenge and other resources like HashiCorp Terraform Tutorials. Be there for them if they have questions or want to bounce ideas off from someone. Cloud Resume Challenge does have a social network to lean on, but sometimes it’s nice to talk to someone familiar.
Remember, “Pipeline solves all woes”. Let’s build a pipeline of new IT engineers in as many ways as possible.
Aquaman Out
Links & Blurbs
https://cloudedvision.substack.com/p/hashicorp-field-cto-weekly-yacp-yet
The previous newsletter I wrote with information about ChatGPT.
https://www.extremetech.com/extreme/115131-learn-how-to-program-a-self-driving-car-stanfords-ai-guru-says-he-can-teach-you-in-seven-weeks
Article describing the original Self-Driving Car class taught by Sebastian Thrun and Peter Norvig
https://karpathy.ai/zero-to-hero.html
Andrej Karpathy’s crash course on neural network programming.
https://www.engadget.com/chatgpt-passes-graduate-law-business-exams-210146640.html
ChatGPT is passing professional exams and more.
https://www.darkreading.com/threat-intelligence/chatgpt-could-create-polymorphic-malware-researchers-warn
The folks at CyberArk convince ChatGPT to write undetectable malware. Naughty.
https://singularityhub.com/2023/01/26/deepfakes-faces-created-by-ai-now-look-more-real-than-genuine-photos/
AI is generating deepfakes, or realistic (but not real) images of folks that fools humans determining what’s AI or real.
https://castlebridge.ie/insights/llms-and-the-enshittening-of-knowledge/
Critique on how ChatGPT is creating realistic sounding references that are complete BS.
https://www.netflix.com/title/8123197
Netflix’s Wednesday show focusing on Wednesday from the Addams Family. It’s sort of a season long Scooby Doo episode for goth-type folks.
https://www.youtube.com/watch?v=VVc5QiMedno&t=363s
“Transformed! A Musical DevOps Journey with Forrest Brazeal”, but starting off with the first two songs.
https://cloudresumechallenge.dev/
Forrest Brazeals Cloud Resume Challenge which gives folks a crash course on application delivery in the cloud with specific focus on getting certified in your cloud of choice.
https://educationdata.org/number-of-college-graduates
Demographic information on US college grads