1000X

When I worked at Expensify, I remember an exciting period for the company which followed the announcement that we were going to run a Superbowl ad. From my estimates, nearly 15 million dollars were spent on the endeavor, which involved hiring a top-rated advertising agency, 2Chainz (the rapper) and months of preparation for the influx of customers leadership anticipated would surge through the floodgates of Expensify.

Every initiative was dubbed “10X” because they anticipated the company was going to increase its customer base by 10X, and they wanted to be prepared for this huge surge of new paying customers.

They fired me before the Superbowl, and I wasn’t there to see the dissapointment from the team when the ’10X’ goal wasn’t met. In fact, they didn’t even come close to a noticeable increase in business after investing millions into the ad – and 2Chainz wallet.

Before getting canned, I remember the long hours that were spent by the company because they believed 10X was possible, probably and just around the corner. All of the employees with stock options seemed to mentally calculate just how rich they were going to get once the ad dropped.

At that time, the idea of 10X’ing something seemed like an ambitious goal, and it was. However, the pipe dream of thinking that a rapper would somehow inspire accountants to use a new expense manageent tool didn’t get much criticsm of sanity checks. As a result, the ad flopped and a lot of money was flushed down the drain.

Today, their stock is trading at $2.50 after an intial IPO debut of $38…

What I learned from that lesson is that many organizations have hopes and aspirations about what ‘could’ happen, and are more than happy to work their teams to the bone in the hopes that their wishes and dreams come true. In the case of Expensify, I know a lot of people who lost their lunch (as well as retirement plans) because they blindly followed promises from a management team that put their trust in 2Chainz, rather than well-based data about what their customers actually wanted.

Spoiler alert: the financial/accounting industry wasn’t exactly thrilled to receive a miniature-manifesto from the CEO, telling them in an email if they ‘didn’t vote for Biden – they weren’t voting for democracy’. Following that email, they had customer churn like they couldn’t believe; similar to the way Titanic struck an iceberg.

After spending three years in the ranks of VC-backed startups and airpods, I eventually landed back in my Aeron chair as my own role as founder/CEO of my beloved agency, Tripleskinny. Instead of dealing with 50,000 customers, I was happy to solely focus on a smaller portfolio of clients and deliver results for them that they could count on in the world of design, development and marketing.

As time went by, I began to experiment with AI and went from being a casual user to somebody well-versed in the capabilities, complexities and modalities of artifical intelligence. In a matter of years, it stopped being a tool that I used, and became ‘the’ tool that I used (and continue to) to run Tripleskinny.

Using AI is a lot like spearfishing. You take a deep breath and then go ‘underwater’ for extended periods of time to hunt for your dinner; patiently waiting, watching and learning. Eventually, you have to come back to the surface for another breath before going down to spear the fish you had your eye on during your first few dives.

For the last few years, I’ve operated in this capacity; going ‘under’ for an extended period of time to observe and learn about the proverbial undersea world (in this case, AI) before returning to the surface to share my findings with those around me.

A few months ago, I watched a video where a top engineer at Google talked about the efficiency of his engineers when using AI; going from 10X, to 100X and now working at roughly 1000X the efficiency in their daily output.

To some, this seems like a far reach. However, I’d like to help deconstruct what that looks like in a practical way.

A well-trained coder can write between 10-50 lines of polished code in a single day.
A well-trained AI can write up to 10,000 (at the time of this writing) lines of code per day.

Compare this with:

A well-trained plumber can fix between 3-6 toilets per day.
A well-trained plumber using AI can fix between 3-6 toilets per day.

To those who believe that blue-collar workers are the next batch of millionaires, I’d highly suggest they revist this notion. Not all work is capable of benefitting from the AI revolution taking place in the industry now.

Yesterday, I had a small mountain of work to complete, which included:

• Backup 30+ website environments, run plugin/core file updates on each installation, test thoroughly, push staging environments to prodcution and development environments.
• Audit and improve website copy on 50+ pages
• Create 3 landing pages
• Advanced setup & integration of GA4 (analytics), Google Tag manager, set up conversion tracking, build funnel reports
• Google Ad Account Setup, create 5 ad campaigns with highly advanced targeting metrics, keyword targeting, negative keyword values, advanced bid strategy
• Develop 2 custom plugins to take website form entries and write them to PDF documents, storing locally and sending customers a copy
• Write 3 instructional course lessons
• Audit and test 50+ page website for functionality and improve UX/UI, design and customer journey
• Register EIN, link payment processor and submit tax documentation
• Write terms & conditions, privacy policy for website
• Develop custom referral calculator to show referral revenue
• Update 50+ website pages with properly formatted graphics, animated backgrounds and interactive user elements
• Develop onboarding form for course, then personalize every element of the course based on the user intake form
• Fix and optimize checkout experience for website
• Update website with 50+ images from client
• Create 6-touch email and SMS outreach program
• Fix broken lead routing
• Meet with client to discuss new website project
• Generate 51 product mockup photos
• SEO optimize 50+ pages on website
• Update each product page to show mockup in hero background
• Migrate website between servers, update DNS records
• Set up new Google Workspace for email account and configure SPF/DKIM records
• 2 loads of laundry – fold and hang
• Redesign 2 website homepages and create full audit of improvements to make on the V2 build
• Clean & organize pantry
• Vacuum & shampoo carpets
• Wash & clean car
• Clean garage
• Drop off 43x at the store for a new trigger installation

Every item was completed by 9pm, and I still had time to grab dinner with a friend and spend an hour or two catching up on life. When all of it was said and done, I calculated that I spent less than 5 hours in front of my computer.

At the end of my work day, I asked Claude how many ‘human hours’ were spent accomplishing what had been done. The answer was staggering.

Nearly 1,000 human hours and between 8-12 weeks of ‘agency time’ that would have come at a nearly six-figure cost were humans to have executed every task on billable hours.

How do you 1000X? That’s how.

What has been my learning in all of this? There’s a changing of the guard as it comes to the world of technlogy and the way people work. When I speak to others about the nature in which AI is changing the landscape of work, it comes from a position of having boots on the ground in the very battlefield where the change is happening.

In every way, I hope that I can share these findings with others, because I have firsthand seen the way in which using AI has changed my life and transformed me into a present Dad when I’m with Atlas, rather than one who is perpetually behind and feeling anxiety from a to-do list that seems inconquerable.

The future isn’t coming. It’s already here.