The AI Strategist - 07.07.23
Designing & Building AI Products & Services
In the news - Today I start an 8 week course from MIT. I can’t wait to get stuck in!
Why this matters - my role at CHAPTR requires me to bridge business and technology strategy. I want to keep my technical skills sharp. It is pretty obvious that most organisations are contemplating their strategy for Generative AI. A big challenge is considering how smaller ML models can diffuse knowledge and understanding into foundation models - this fusion will power AI products and workflows of the future. This course will ignite a few new ideas in this space.
2. Is AI Opening the Door for the 100x Founder?
In the news - An interesting VC perspective on AI startup productivity.
Why this matters - the most interesting quote:
“We often hear the term the 10x developer; the almost mythical elite engineer who's 10 times better at their job than anyone else. But through AI, we have to ask, are we now entering the era of the 100x founder?”
I spent 3 years seeing Faculty go from 60 to ~250 people and 1 year at CHAPTR going from 5 to 30-odd people (the people are lovely, they are not odd). Scaling processes is really hard. A huge part of this is trust and explainability. Seeking truth and critical thinking is imperative. We aren’t yet at a stage where we can put our faith in outsourcing tasks to AI that are crucial to our growth and strategic direction. That said, I absolutely do believe in a future where founders have an opportunity to build and reimagine AI workflows and automations to grow. This will be a key source of low-OPEX growth and competitive advantage. I am learning first hand that it isn’t easy to get right.
3. Should You Start a Generative AI Company?
In the news - Headline quote:
I am thinking of starting a company that employs generative AI but I am not sure whether to do it. It seems so easy to get off the ground. But if it is so easy for me, won’t it be easy for others too?
Why this matters - Two reasons …
1 - I love the simplicity of from Scripted to Generative. Making the argument that there is a continuum of 100% human to 100% foundational model driven. At CHAPTR we call this 'embracing the blend'.’ We are talking about the blend of human experts (i.e. authors) with technical experts (i.e. AI engineers) to create something amazing (i.e. our first B2C product). The startups that win (sustainable competitive advantage) will nail exactly where they want to be on this spectrum and communicate it to customers through inherently safe, trusted, life-enhancing products.
2 - I laughed wryly at the naive assumption that ‘it seems so easy to get off the ground.’ This couldn’t be further from the truth. Try tackling hiring, product market fit, the pace of foundation model development, UX, AI Safety, AI ethics, AI regulation! We are one year in and making great progress. But … almost all of the challenges of being a Generative AI Startup are non-trivial.
4. The Rise of the AI Engineer
In the news - We are observing a once in a generation “shift right” of applied AI, fuelled by the emergent capabilities and open source/API availability of Foundation Models.
Why this matters - This blog perfectly encapsulates the challenges I have faced over the last 12 months in going from Faculty to CHAPTR. It is a bit of a quantum leap where previous thinking and methods (traditional ML) need to be questioned given that we are now in an ‘LLM-enabled AI’ era. It has implications for:
1 -Corporates. If there are only 5000 LLM researches in the world … how do you hire one? How do you get value from them?
2 - Investors. What happens to the pre-LLM startups of 2015-2022? How do you find the next Inflection AI from the class of 2020 onward (i.e. a natively GPT-3 business)?
3 - Teams. How do this new blend of (ML engineers, Full stack engineers, Data Scientists, AI engineers and all the other commercial people) co-exist, in harmony, to create value?
4 - CFO. That 6 month project (traditional ML) with a fairly decent ROI just became 6 x 1 month POCs (‘LLM-enabled AI’) with much more risk. Do you roll the dice or play it safe and avoid Generative AI investments?
5. Software 3.0
In the news - OK, this is a blog from April … but …
Why this matters - If you are struggling to understand what AI software was all about before Chat GPT and where to start right now, this is a good place to start. It builds on this well-known blog. Another great source here to help you break down the definitions.
6. Changing the business model due to Generative AI
In the news - Totally fascinating example of an AI business that is being disrupted by Generative AI.
Why this matters - I think that once the Generative AI hype dies down, the amazing businesses of the future will be proprietary AI + foundation models = great products. Customers don’t care what happens under the hood but obviously businesses need to invest to differentiate and shout about it. Phrasee is aiming to differentiate through control. Time will tell. Getting this right is about survival. If you want to go a bit deeper on the intersection of recommendation engines and Generative AI check this out.