Milan, co-founder, is taking a position with regards to rising AI tides
Dec 10, 2023
Hey there 👋🏻,
Milan takes a position with regards to rising AI tides. ✍🏼
AI won't take your job. It's somebody using AI that will take your job.
Historically, AI developed in waves of breakthroughs, inflated expectations, and disappointments. Remember IBM’s Deep Blue from the ‘90 that won a chess game against the world champion? It made quite a hype, not many of you remember now. Remember also Watson, the hype of 2011 that could beat quiz champions? Maybe not.
There is something qualitatively different about the paradigm-shift that was created by ChatGPT. Everybody can use it, and finally get their own first-hand experience with what can be expected from AI. ChatGPT has raised the expectations of users who now legitimately expect AI everywhere. There is no coming back to the pre-ChatGPT era.
OpenAI has essentially lifted AI from a nice-to-have passion of computer science researchers to a naturally expected feature of any software system. And that is huge!
"AI won't take your job. It's somebody using AI that will take your job." ~ Richard Baldwin, professor of Economics
What does it mean for B2B sales? A lot! We’ve been interviewing B2B sales professionals to understand their workflow and identify opportunities for AI to transform their jobs. We are grateful to all of you who participated in this research, and if you haven’t yet, we’re happy to schedule a call and learn about your practices. We are also taking a deep dive with first customers in order to build an AI product that will transform the everyday life of B2B sales people.
What we see is that generic chatbots are not really made to relieve sales professionals from time-consuming information gathering tasks. ChatGPT does not really know who sells {what, how, where, to whom} and who needs {what, when, why} to {produce, distribute, sell} what.
Because of this, the support one can expect in the sales workflow is currently quite limited. We are building something specifically tailored to bend the whole sales-related information space into a few efficient interactions that can put B2B sales efficiency on steroids.
Techno-optimists
In the current AI tide, inflating both expectations and fears, we take a clear position as techno-optimists (cf. Marc Adreessen’s Techno-Optimist Manifesto).
In addition to reasons given by Marc Adreessen, we are optimists also because the human is in the center of our R&D. We are building technology that extends the human intelligence, and hand-in-hand with the human pushes the boundaries of previously possible.
This is especially important in the context of growing evidence that the average human IQ in developed countries might be seeing end of growth or even decline.
Augmenting (rather than replacing) the Human
Our view on technology as an extension of the human intellect, is in the continuity of some of the fundamental positions in Computer Science that have shaped the modern, commercially available, information technologies. Firstly, we refer to the Memex – a concept proposed by Vannevar Bush in his article "As We May Think" (1945) that inspired personal computers, the Internet and the Web.
Secondly, we refer to Doug Engelbart work on augmenting human intellect – NLS Augment (1969) that set the ground for modern human-computer interfaces and collaborative software.
At its current frontier, the way how humans and machines can achieve complex tasks together involves AI. We explore ways how AI can support the human in navigating complex information spaces, making predictions, learning, and making decisions. This research involves both the work on core AI systems and the work on novel ways to interact with them.
Building for Digital Natives
The human is now almost inseparable from technology. The human brain, during its formative years, is heavily exposed to technology that is involved in many cognitive tasks. We are only beginning to understand the implications of this phenomenon, but the following is clear: the expectations from technology are changing, digital natives are a growing part of the population (and already 38% of the workforce), and they have specific needs.
We are building for them.
There is more to AI than chatbots and current LLMs
The popularity of Chat GPT has established a strong, almost synonymous, association between AI and a conversational interface. Conversational interfaces have important limitations and are known to increase the users’ cognitive load more than classic graphic user interfaces. As other AI researchers, we are skeptical about their widespread use in the future. We are interested in going beyond.
Independently of the form of interaction being conversational or not, we can also expect much more from an AI system when it comes to supporting the human in tasks that require accurate and timely information. This is precisely what the context of professional interactions and sales is all about. This view is consistent with Yann Le Cun’s view : “There will be better systems that are factual, non toxic, and controllable. They just won't be auto-regressive LLMs.”
Rather than focusing on one AI feature to add to a software, we are interested in a future where AI can manifest itself throughout the whole software system, supporting the user in multiple tasks, in many different ways.
We are aware that in the current world, these positions, however reasoned and moderate, may appear radical. It is maybe how it should be.
