A vision of the future of software

Sometimes one is given a glimpse of the future that is so compelling, so unshakeable, that one has to turn their life around for it. This is that glimpse for me.

TLDR:

• 1. to create value, make a unit of time less painful/more joyous, or decrease pain units of time
• 2. ‘saving time’ is essentially computational reduction - do unpleasant things in fewer steps
• 3. Discovery for ‘time-saver’ products is an incredibly manual process
• 4. Guava: the system of record for workflows, to identify then build automations that objectively save time
• 5. Yes, it’s a moonshot. But if it works…

1. to create value, make a unit of time less painful/more joyous, or decrease pain units of time

My first real obsession was the question, ‘what drives human value?’ I wanted an explanation to cover innovation and consolidation, wealth and poverty, both Chanel handbags and HR software. I never settled on an exhaustive framework, but I accepted a simple paradigm: time is a core axis, which is then qualified by joy or pain. ‘Human value’ is then simply defined as the sum of the area ‘under the curve,’ or a fancy $\sum joy(t)*dt - \sum pain(t)*dt$

As simple as this sounds, this paradigm gives a crystal clear direction to creating value: make each unit of time less painful/more joyous, or make joy longer and pain shorter. This will always be true. And while it’s impossible to quantify joy/pain, time can be measured (okay yes, actually what matters is the perception of time, but pain makes time feel longer, so bear with me for now). Cue Mr. Bezos:

What's not going to change in the next 10 years?' […] In our retail business, we know that customers want […] fast delivery. […] It's impossible to imagine a future 10 years from now where a customer comes up and says, 'Jeff I love Amazon; I just wish you'd deliver a little more slowly.' Impossible. And so the effort we put into those things, spinning those things up, we know the energy we put into it today will still be paying off dividends for our customers 10 years from now.’ - Jeff Bezos

2. ‘saving time’ is essentially computational reduction - do unpleasant things in fewer steps

While it might not explain the value of luxury handbags or social media, our time as joy/pain framework allows us to confidently delineate the world of software that pertains to ‘making pain shorter.’ This is effectively computational reduction: what used to take X steps, can now be done in <X steps.

This is typically done by choice-less steps while keeping choiceful ones. This is a fancy way of saying ‘automating the busywork:’ the parts that people don’t enjoy, being compressed by software. Excel in the 80s: instead of manual calculations, save time with auto-calculations. Expedia: fewer steps to compare flight options. Zapier: fewer steps to transfer data across tools with stock integrations.

(For further reading on computational reducibility, I heartily recommend Stephen Wolfram’s essay on a possible unifying theory of physics, but be warned, it’s roughly 25,000 words or ~100 pages.) Or start with his Ted Talks.

3. Discovery for ‘time-saver’ products is an incredibly manual process

But how did these time-saving tools come about? Historically, the process of this reduction was investigative, a creative and iterative approach by founders who would experience or ask their way through some problem space. It relied on curiosity, insight, agency, and high drive. Then, the founder had to bet on it, dedicating costly time and resources to build a solution. And for each product that succeeded, tens failed. It’s incredibly hard to preemptively validate - especially at a quantitative level - the value of a certain reduction, without the sunk cost of building the product and testing some go-to-market motion that must be cheaper than the product development.

Recent AI tools have drastically decreased the cost of development. This cheapening will inevitably lead to more attempts at computational reduction, since the same amount of time and money will now yield more ‘at bats’ for testing products. But cheaper development also means that smaller solutions are now feasible as well, since a small customer solution won’t require a whole team of devs anymore. But even cheaper development doesn’t solve the problem of discovery - there will still be wasteful development as people invest in solutions that don’t always pan out.

Because that’s not the end game. The end game requires a double layer: data ingestion and analysis.

4. Guava: the system of record for workflows, to identify then build automations that objectively save time

Currently, there is no system of record for most people’s work processes. If someone is taking 10 steps to do something that software could reduce to 7, the only way to discover that inefficiency is to have a curious would-be founder stumble upon the right questions with the right people. Within siloed applications, product managers use Pendo or Mixpanel to track their users, but any workflow that goes across apps (or through untracked apps) is lost forever, captured only in the lossy memory of that worker. Where is the Pendo for your workflow?

There is no way for people to proactively surface their workflows, beyond a standard complaining about the parts of their jobs that they consider ‘annoying’ or ‘frustrating’ (read, painful). But this is an incredibly manual process, hamstringed by the fact that the people complaining are not always the most observant, the best incentivized, or the most informed.

A general solution here needs to follow human work, so that the inefficient workflows that used to be identified only by observant users and shrewd founders, should be tracked somewhere. Our moonshot bet is that this data layer can and needs to be built, exclusively based off the visuals of someone’s screen - the only place where all the work intersects at some point in time.

But once this data layer exists, it becomes possible to accurately capture and map workflows, across jobs and industries. And here is where our second moonshot comes in: finding the areas, big or small, where steps can be reduced. This process currently happens at the micro level with founders going on a limb for product design, or with consulting companies taking massive paychecks to ‘streamline’ processes. In an era of expensive software development, only large pockets could be reduced. However, with cheaper and faster software, coupled with the proliferation of APIs across tools, pockets of any size can be reduced - they just need to be found.

5. Yes, it’s a moonshot. But if it works…

This is the moonshot of Guava. We don’t just want to streamline the workflows of one vertical or industry: discovery in every instance shouldn’t just be something for early-stage startups, it should be constant, pervasive, accessible. As AI has decreased the cost of development, we expect and have already started to see software solutions for smaller and smaller markets.

We have this unshakeable vision - one of people proactively putting forth their workflows in order to receive automations, a thousand different versions of custom software at the biggest and smallest scales, spun up by owner-developers who can match the effort to the opportunity. Risk-taking founders will always be needed to take a swing, but those swings will be of all sizes, and lead to ever-higher batting averages.

This would entirely change how new software is brought into the world, from inception to execution. Any individual, team, or organization would be able to continuously streamline human time in order to eliminate, at a system level, any bits of ‘busywork,’ of any size. Which is fitting, because that is our mission: to decrease the amount of busywork in the world to the absolute minimum. Because we just want to make painful time as short as absolutely possible.