Checkpoint in Q&A format. Feel free to respond on LinkedIn if you have questions. Questions were generated by surveying some colleagues who've been supporting me.
Snark and condescension in the questions added for effect. My friends are very kind.
Are you working on “starting something”?
Yes.
What’s your business model?
That’s a valid question when someone’s building a business. I’ve made the mistake a while ago of conflating building a business and designing it. They are two very different phases in the process of creation.
At this point, I’m validating the feasibility of a thesis. Market and product research and development with hopes of understanding the feasibility of creating a product experience which could drive a scalable business.
In short, it’s too soon to be concerned with the model because I’m not clear if the product is possible without a more major R&D investment.
What are you trying to learn?
I'll dig into this question later and touch on how I got here, but it's "Can my experience managing, building, and deploying teams of humans to produce high-quality work inform how AI systems should be designed." Or, in other words, how much does "environment", "leadership", "trust" and "organization structure" contribute to the work outcomes for AI-first systems.
What’s your process?
Great Scott! Don't ask a reformed Innovation Consultant this. The difficulty of any creative process stems from the intuition of when to pivot, skip or exit.
Market Discovery, Ideation & Modeling, Product Prototyping, Validation, then Market Launch. After Market Launch, I’d enter the product development and GTM loops to find all the first fits between the Product, Market, Channel, and Business Model.
I’m crafting an experience that requires some capabilities beyond what I’ve seen from the best AI-centered products in the market and definitely beyond the out-of-the-box abilities of the open and closed models I’ve used thus far.
Wait, you’re trying to out-AI Google and OpenAI?
No. I’m not going to innovate in AI, per say. I’m seeing something many in the space aren’t (or haven’t executed on publicly).
My perspective is unique. I’m a deeply technical entrepreneur with experience growing a tech-enabled software development agency at Gigster, building an innovation group in V1., and working on two very different distributed processing systems where a billion+ messages per day were processed in seconds and available for insights at Rollbar and AddThis. The work I’m doing is a culmination of this experience considering what the new AI-tech can bring, and how it should integrate into software.
Therefore, I’m taking my perspective with the hope to out-product-design and out-software-engineer Google and OpenAI to create an experience they may not have envisioned, yet. Let’s remember this is why startups exist. Facebook wasn’t the first social network, Slack wasn’t the first collaboration software -- they innovated in the experience of a known product experience. Eventually, they invested in R&D to move forward the state-of-the-art, but that wasn’t what made them special out the gate. We see an experience that should exist or a bad experience that shouldn’t and fix it. The hope is the experience which solves a problem for the customer is transcendent enough to disrupt existing solutions.
Won’t They just copy you if it works?
Based on my work to-date and project solution architecture -- it’s sufficiently complex that anyone who appreciates the direction and traction would want to partner vs try to build it on their own. The system architecture, design and, eventually, usage data can act as a solid moat. I don’t see any open source framework making it “easy” to accomplish what I’m going for, yet, but that landscape evolves rapidly. I'm also learning first-hand how many insights one must have to get this experience right.
Why aren’t you talking about your offering publicly?
Simply, it's not an offering, yet and I won't sell an experience I don't clearly believe is possible to deliver.
I am sharing what I’m learning about the space, market, forecasts, predictions and the “inner-workings” of the state-of-the-art updates in the AI community. That helps me synthesize and allows me to have thoughtful discussions on the periphery that help guide me.
At the moment, the core insights I’m building on require validation.
Many of the ideas are in a fragile state where they’re morphing and changing on a daily and weekly basis as I learn. Given this reality, the “why”, “what’s different”, “advantages” and positioning are all in too much flux to have a conversation which won’t feel schizophrenic from the outside in.
Most would confuse this essential chaotic stage of creation as a “lack of direction” or worse “as a failure of process.” For those familiar with early-stage exploration, I’m in a normal place where my revelation from last week is less interesting after I learn new facts. All that being said, I can’t wait to leave this phase as it’s psychologically the hardest on me.
Do you have a name for whatever you’re working on?
Yes --SendThis.
What have you been doing the past quarter?
Beyond spending a third (or more) of my time just reading and synthesizing what’s happening in the market and research scene, I’ve been prototyping and building.
I’ve built an end-to-end prototype that is relatively high-fidelity. Initially focused on system design, architecture, cost and testibility. Now I'm focused on the AI-tech.
Sadly, when I user-acceptance-test and integration test the system, it’s not hitting the mark on the level of quality needed to meet my expectations. I’ll take a pause soon and try to write up some learnings.
Sharing what I have now would demonstrate that I can build software, and that I’ve been working hard. It wouldn’t demonstrate that I have built a competitive experience to what’s out there.
At this stage in my career, I don’t feel the need to prove to myself that I can get “something” out. I’m a little past the “build a portfolio” stage and want to build something that either “wows” consumers or “inspires” the small subset of the tech community that doesn’t revel in betting against people.
Given where I am, and the progress, I do feel that I have a few more rocks to turn over before I call this problem intractable given my resources and expertise. My thesis is continually guiding me in a direction that aligns with my skills and unique personal advantage.
What’s the product you’re envisioning?
SendThis is an AI-first Agency for SMBs and creators. Most can’t afford a personal assistant, analyst, or writer and their business model won’t allow for it. They’d benefit from those services. I believe it’s now possible to outperform the lower-cost knowledge workers -- this is part of the broader thesis my work should validate.
The hope is to feel like the customer is working with a partner agency. There are plenty of tools boiling the ocean with chat but what will make SendThis different is the quality bar. In the end, it'll be better because of the fact that I'm operationalizing and systematizing more of what matters to facilitate great work outcomes. What I just said is the key assumption I want to validate -- "Can my experience managing, building, and deploying teams of humans inform how AI systems should be designed."
What’s different about it vs Gemini/ChatGPT/Anthropic?
TLDR; too much technologically.
At the core, there will be a lot of products and they’ll need to decide on the experience the customer expects while creating the work product with the tool.
Does a customer want a hands-off experience or do they want to drive? For example, does the customer hope to act as an art-director with a team of AI artists, or are they a business owner who wants a director to achieve some business goal.
When I look at the AI consumer landscape, I see that most tools and frameworks are focused on customer-directed synchronous collaborative work [excuse the mouthful] with AI systems. More asynchronous tasks such as “deep thinking” or “creative” projects haven’t been getting as much attention or, potentially, traction to rise to my radar. The underlying assumption is that humans directing the work will produce a better result and have a strong idea of how they want to achieve the work. The assumption is that the “human user knows what’s best.”
How do you expect to compete against billion dollar companies?
Let’s make sure we don’t conflate
(a) “building a competitive business”
with
(b) “defining a point of view and validating that it’s technically feasible.”
The latter (b), as any Innovation Leader can tell you, is very hard with groups of people. I didn’t start with co-founders. I have learned that it's not prudent to “recruit” folks just because we all want to start “something” and have some personal chemistry.
What’s important is the insights the project, and eventually organization is founded on. I’m working to build a thesis that makes sense in the highly volatile landscape. Anything worth building that is disruptive must, by its nature, do something better and cheaper than before. It’s a non-starter to release something that doesn’t have some step-change advantage.
At the moment I need to build something that inspires me to double down in a direction. I don’t need to out-AI all these scientists to win, and don’t think that any company starting should have that be a requirement for success. My hope is to “out-software” and “out-product” these scientists who are more focused on publishing papers than customer needs. That’s a much more tractable goal.
Doubling down means finding partnership in a team and, potentially, capital depending on the needs of my early partners.
I am grateful to have resources to finance my exploration. I enjoy the optionality to change directions without managing that change with any other stakeholders. That’s something that, once you have others, isn’t reasonable to expect.
What’s winning to you?
I recognize that from where I am, dreaming of building a massive, earth-shaking organization isn’t productive. I like to set goals that are a little more reasonable.
There are at least two different winning paths which are reasonable from where I’m standing:
1. I build something I love that I’d use. This win leads to me doubling down.
2. I fail to build something I love in a reasonable time and discover a use-case that’s adjacent that I can pivot.
On a daily and weekly basis, I’m enjoying the incredible challenge and learning that comes from doing this type of creative work.
How close are you to getting a product that you can validate yourself and with others?
My goal in January ‘24 was to get something in testers’ hands the first week of March. I’ve missed that target.
Although, I am proud of the progress given how little I knew when I set the goal -- I want to re-establish guardrails.
At the moment, I will set a new schedule after I finish some of my current tests with the effectiveness, cost and work required to fine-tune models.
Have you selected a Cloud Provider?
Yes, AWS. Why? I adhere to a greedy procurement algorithm and sorted options alphabetically.
Why do you share on LinkedIn?
I wrote a post about that. Read more
Do you think you should see someone about your conversations with yourself?
Yes. Feel free to recommend a professional.
