Welcome to the third edition of Barefoot Bytes. I’m writing today from the largest log cabin in the country. Sounds peaceful, until you learn that there are 18 adults and 13 children running around this madhouse. I digress, let’s dig in.

Learning is Broken

The rapid pace of change is making it really hard to keep up. Even for professionals like myself. A personal story: in 2001, I dropped out of the Computer Science program in college when I realized they had just a single course on web development, opting instead to teach myself programming from books checked out at the library while earning a liberal arts degree. The curriculum just hadn’t caught up with what was happening on the internet.

More recently, I sought to improve my prompt engineering skills, which are critical when interacting with LLMs and other natural language AI tools. I jumped into Udemy to find a course. Looked at a few, checked ratings and recency, picked a course and got started.

Lesson 1: I fire up my Jupyter Notebook, sip coffee, crack knuckles, and start writing some code. Line 3:

What? I get an error that this function is no longer supported. The course was published two months ago. In that time span OpenAI released 33 newer versions of their library and this course was completely obsolete.

Even source documentation is unreliable. I’ve been hearing this from my engineers a lot:

Universities, books, online courses, and even source documentation cannot teach us how to use these new technologies. So how is one to learn these days?

I believe that it’s time to make a deliberate switch from passive to active learning. Learn by doing. Rather than reading a book, or taking a course, we must engage with the technology directly. To learn we must experiment, tinker, and build.

Further, the rapid pace of change means the developer community is more important than ever. As we build and experiment, your own knowledge becomes the support system for others. We see this in thriving open source projects, GitHub comments, and Stack Overflow.

These communities are the new classroom.

Regardless, at a macro level, we are going to need to completely rework how both formal and informal education works. This is especially important because the rate of change is increasing dramatically. Mary Meeker published the following chart in 2019:

What does it look like today?

SaaS is Dead. Long Live SaaS!

There has been a lot of chatter about the future of Software-as-a-Service in the tech landscape, with some claiming that the SaaS model is dead. The reasons you hear for this vary, but here are a couple that I’ve heard recently:

• Consumers are tired of paying for subscriptions. 37Signals put out a manifesto last year launching “Once”, a new product line where you can purchase their products with a single lifetime license.  I do think that the pendulum has swung pretty far towards SaaS and this old-now-new model might be making a comeback. The market will find a new equilibrium.
• Compliance is only getting harder. Large enterprises are losing trust in third party platforms and bringing more in house. Dealing with mounting privacy regulations across the globe, they are doing this to mitigate risk.

OK sure, I do believe that there is some truth to both points, but they are not SaaS killers. Salesforce is not going out of business any time soon. The incumbents will be around for a long time.

What I do think will need a massive overhaul however are SaaS startups. The old playbook started with an idea – a gap in the market. Put together a nice pitch deck, raise some seed money, and go heads down building product for a year under the radar. Launch your MVP, raise your Series A, and you’re off to the races. Now you have what is called the “First Mover” competitive advantage. And there’s the rub. It no longer takes a competitor a year and a lot of funding to build an MVP. And the reason is generative AI. Maybe not today, but soon these tools like Devin are going to rapidly accelerate the process for building software to the point that it destroys any first mover advantage. This is already widely known in the investor community.

So what is the new competitive advantage? I believe it will be proprietary data. When an investor asks why a concept can’t easily and quickly be copied, the answer will be that the startup has access to some unique and proprietary dataset that enables the training of models that others can’t. If I were looking for a new startup concept, I would be looking at existing businesses with interesting datasets. I would come up with a concept that is non-competitive and potentially complementary to their business, but needs their data. Either pitch them to invest, or sign an exclusive licensing agreement and head to Sand Hill Road to raise some capital.

All of that is to say, the time and cost of software development just started a race to the bottom. Meanwhile, the value of data has increased exponentially with the release of these new tools.

Bytes

• Nvidia is now the world’s third most valuable company, leaving only Microsoft and Apple with larger market caps. Their vision is so much more than chips. If they execute, they’ll be number one by 2025.
• I’ve been using Fireflies.ai for meeting note taking and have been very happy with it. Allows me to stay present without the need for a lot of manual note taking. Also, if you want to make a real power move, you can not show up to a meeting and just send your notetaker instead.
• I have also been using reclaim.ai to help protect my time. It protects 2 uninterrupted hours in my workday, and blocks off 15 minute cool-down time between meetings. If you’re juggling a tricky calendar where others can just book you, I recommend it.

How can we can help your business?
Email us: [email protected]

Share this article