Secret cyborgs, the typewriter and how AI crosses the chasm
“If a machine is expected to be infallible, it cannot also be intelligent.” - Alan Turing
AI, the typewriter, and Roger’s bell curve
So, “secret cyborgs” are a thing... boy oh boy what interesting times we live in. I became familiar with this term a few days ago with Ethan Mollick’s newsletter “Detecting the Secret Cyborgs” here on Substack (by the way, if you don’t already, definitely give Ethan a follow, his work is fantastic). After re-reading his thoughts on the phenomenon, I have some observations of my own to lay out here, but before doing so, let’s explore the term “secret cyborg”. Ethan broadly defines them as: employees who use AI, specifically Large Language Models (LLMs), to boost their productivity without revealing their use of such technology to their employers.
The core of Ethan’s argument is that AI makes for an excellent individual productivity booster, but fails to deliver true organizational value due to reasons he lays out more thoroughly in another article: AI is not good software. It is pretty good people. And I mostly agree with him here - it’s one of the prime problems with the current AI bubble - lots of products, lots of individual value, and relatively little impact on the organizational bottom line.
Thanks for reading Future of Product! Subscribe for free to receive new posts and support my work.
Ethan posits that the reason why so many employees are taking the extra step of concealing their AI usage largely comes down to three main pressures that they experience in the workplace:
Organizational policies that ban AI usage obviously give employees an imperative need to hide their use of AI tools.
The perceived advantage of AI-generated content appearing human. Or the perception that by producing at a superhuman level without their employer’s knowledge, employees can get an edge in promotion opportunities.
Fear of job redundancy due to AI automation.
None of these are especially surprising, but what they tell us about the potential future of the AI industry is. Towards the end of his article, Ethan drives at a valuable point - the people who are best positioned to create revolutionary innovation via AI technologies aren’t the managers, consultants, or IT departments that usually lead software adoption, but rather the individuals who do the jobs themselves. It’s a topic that’s come up on the podcast numerous times - AI solutions are best leveraged by people who feel and understand the pain point that they’re trying to solve.
These insights led me to look into some other technological innovations with similar early challenges in an attempt to better understand how the AI industry can make the jump to providing transformative organizational value. Surprisingly, the answer was a little further back than I expected...
The typewriter existed in varying forms for centuries before it reached widespread adoption. As far back as 1575, some Italian printmakers were utilizing rudimentary versions of the technology to speed up the process of impressing letters onto paper. Despite these early explorations, it wasn’t until the mid 1800s that the typewriter saw a major increase in activity from inventors - with the first commercial model arriving on the scene in 1873, and the famous E. Remington & Sons model hitting the market in 1875. This invention was a revolution (or at least it should have been, but we’ll talk more about that in a second), for context - the average person is capable of writing about 13 words per minute by hand (around 29 per minute for trained calligraphers)... the manual typewriter increased this to about 40 words per minute on average, with trained typists consistently hitting speeds of 120 words per minute or more on the Remington.
But, rather surprisingly, mass adoption of the typewriter wouldn’t come about for another two decades - with the beginning of what has now become known as the ‘progressive era’ of American culture in 1897.
Interestingly enough, that no man’s land of limited adoption has some striking parallels to the trends that Ethan pointed out in AI adoption today. In that roughly 20 year period of sluggish growth, there were in fact many early adopters of typewriter technology, they just happened to be individuals rather than businesses. Some high-profile examples include Friedrich Nietzsche’s adoption of the technology in 1882 to help him remain productive in spite of his worsening blindness, and Henry James’ switch to dictating his stories to an assistant who used a Remington typewriter - a change that today is widely credited with shaping the cadence and style of his later works.
But perhaps the most noteworthy early adopter of the typewriter was also the first person to give the world a small glimpse of its potential. Mark Twain bought his first typewriter in 1874, and had the following to say about it in a letter that he typed to his brother, Orion:
“the machine has several virtues i believe it will print faster than i can write. one may lean back in his chair & work it. it piles an awful stack of words on one page. it dont muss things or scatter ink blots around. of course it saves paper.”
In 1883, the world unwittingly got its first taste of what the typewriter was capable of, with Twain’s memoir “Life on the Mississippi” - the first manuscript ever typed on a typewriter. To exemplify just how ahead of its time the typewriter was (and just how unique Twain was in his adoption of it), it’s helpful and rather mind-boggling to realize that the first practical fountain pen wasn’t invented until 1884 - almost exactly 10 years after the Remington typewriter hit the scene and Twain first bought his.
It would be 14 more years before typewriters were a staple of the American business.
All this begs the question, why? Why would such a transformative technology languish in relative obscurity while the world suffered through ink blots and cramped hands with inferior solutions? The simple answer is that this is just the way that disruptive technology is adopted - very slowly, and then all at once. If we look at the problem through the lens of the technology adoption life cycle model, it’s pretty clear to see that this trajectory isn’t isolated to typewriters and AI.
Originally developed in the 1950s in a sociological study of farmers and their adoption rates of new agricultural technologies, the theory that would go on to become the innovation adoption lifecycle has been demonstrated in both principal and practice time and time again since. The model is structured around Roger’s bell curve:
Innovators - generally wealthier, can afford and are more inclined to risk
Early adopters - seek speed and cost savings, play an active role in leading their community, generally less wealthy
Early majority - active community members, moderate in their risk-taking behavior
Late majority - more conservative in risk-taking behavior and less socially involved
Laggards - eldest of the groups, very conservative in risk-taking behavior
Now, obviously, the intersection of business and technology are more complicated than any model could account for and highly dependent upon unpredictable societal factors. But, the trajectory of nearly all of the most disruptive technological innovations of the past several centuries can more or less be seen to align with the projections of this model.
It seems that we are, with AI, in a somewhat similar position to the typewriter in the mid-to-late 1800s. A smattering of individual innovators and early adopters are making the most of the technology, while businesses lag behind. It's not hard to imagine a future where AI, like the typewriter, becomes an indispensable part of every business operation. But we're not quite there yet, and for a lot of the same reasons that the typewriter took so much time to infiltrate the business world.
Firstly, businesses were, have always been, and remain, inherently risk-averse. Adopting new technology can be costly, both in terms of the financial outlay and the time and resources required to train employees to use it effectively. The early models of the typewriter were expensive, and largely seen as a luxury and a liability, just as many AI technologies are today.
Secondly, there's a lack of understanding about the potential benefits of the technology. Just as businesses couldn't envision the time-saving and efficiency benefits of a typewriter back in the 19th century, many organizations today fail to see how AI can enhance their operations, productivity, and ultimately their bottom line.
This is where Geoffrey Moore's Crossing the Chasm theory comes into play. Moore's theory posits that there’s a "chasm" between the innovators and early adopters of a product and the early majority - who represent mainstream success in the context of tech company growth. This chasm represents a period of uncertainty and risk, where a product must prove its worth in order to become mainstream. The challenge faced by AI technologies today is precisely this - crossing the chasm. The "secret cyborgs" are those early adopters who have seen the potential of AI and are using it to their advantage. Now, the task ahead for the AI industry is to convince the early majority of the tangible benefits that AI can bring to their work.
Lastly, we're faced with the fear of job redundancy due to automation. Just as early clerical workers may have feared the typewriter would replace them, many people today worry that AI will make their roles redundant. The advent of the typewriter and its eventual ubiquity offers us a valuable historical precedent for how AI might eventually become a staple in every business. Just as the typewriter ultimately transformed business operations and made tasks easier and more efficient, AI has the potential to do the same. While the typewriter undeniably eliminated jobs in the form of clerical careers, it vastly outweighed those losses by creating an entirely new industry of typists and secretaries, many of whom were women entering the workforce for the first time.
This is an often overlooked but significant societal impact of the typewriter, as it played a crucial role in catalyzing the 'progressive era' and the rise of women's participation in the workplace. Similarly, while AI might indeed replace some current roles, it's highly likely that it will also create a host of new job categories we can't even conceive of yet. These new opportunities, combined with the productivity boost offered by AI, could propel us into a new era of workplace advancement and societal change. As with the typewriter, the journey may be slow and fraught with resistance, but if we can successfully cross this chasm, the potential for widespread AI adoption and its transformative impact on the workplace is enormous.
However, it’s crucial to remember that the adoption of such transformative technologies doesn't happen overnight. It's a gradual process, one that requires time, patience, and a willingness to adapt. Returning briefly to Twain, one can also see the prototype for another AI trend of late - the starry eyed dreamer becoming jaded and cynical of a technology they helped popularize. Not long after publishing the first book written with a typewriter, Mark Twain went on to say the following of it:
“I will now claim — until dispossessed — that I was the first person in the world to apply the typewriter to literature. The early machine was full of caprices, full of defects — devilish ones. It had as many immoralities as the machine of to-day has virtues. After a year or two I found that it was degrading my character, so I thought I would give it to Howells... He took it home to Boston, and my morals began to improve, but his have never recovered.”
That Twain perceived a moral corruption inherent to the typewriter provides quite the parallel to the public statements of someone like Sam Altman who today warns us of the existential threat he believes his innovations could unleash. You can read more of my thoughts on that topic here (short version, don’t buy into the hype): Big tech is pulling up the ladder. But I digress...
In conclusion, the trend of "secret cyborgs" is just the beginning. As AI continues to evolve and become more sophisticated, we can expect to see it becoming an increasingly prominent part of our working lives. The key to navigating this transition successfully is to be open to change, willing to learn, and ready to embrace the potential of AI to transform the way we work. Just as the typewriter eventually became a common sight in every office, AI could soon be just as ubiquitous.
3 AI-powered tools from YC Winter ‘23 to check out
The product - Extend promises to revolutionize data processing by transforming unstructured data into structured, actionable insights. It's equipped with LLM technology for unparalleled accuracy and efficiency, spanning various document types.
The use case - a solution like Extend can help you to turn complex data into actionable insights. The tool's adaptability allows it to fit seamlessly into your unique workflow, regardless of your industry. Extend boosts productivity by shifting the focus from data processing to strategic decision-making and growth, which is the true essence of product management.
The product - Gloo provides a domain-specific context for your data. As a search engine, Gloo employs semantic search APIs and robust AI technology to not only ingest and index your data but also feed it confidently into LLMs like ChatGPT. Its unique features include automatic document summary indexing, intelligent embedding generation, and always-on server-side encryption, all fine-tuned to your unique data needs.
The use case - for product managers, Gloo could be a critical link between your organizational data and your LLMs. Imagine connecting ChatGPT to your knowledge base and enabling it to search, cite, and synthesize answers from your indexed data with high levels of accuracy. Gloo even validates AI-generated responses against your knowledge base for added trustworthiness. Furthermore, Gloo eliminates the technical overhead of setting up embeddings, allowing you to focus on strategic product management tasks while it takes care of the intricate indexing process.
The product - Pyq is a powerful tool that simplifies the integration of AI into your development process. Although in early stages, it stands out because it targets specific engineering tasks that can be optimized with AI and offers simple-to-use APIs that can be integrated within minutes. The service lineup includes building content moderation filters, image captioners, image generators, and sales call transcribers. Beyond these capabilities, Pyq also offers customized, task-specific AI for enterprises, facilitating model training, deployment, observability, and maintenance using proprietary technology and human expertise.
The use case - could be really useful for product managers and developers seeking to inject AI-powered features into their projects quickly and effortlessly. It enables organizations to harness their disparate datasets, convert them into practical training data, and subsequently deploy customized, reliable, and private models onto their existing cloud infrastructure. While a consultation call is needed to get started with the main product due to their early stage, Pyq also offers a playground called the "zoo" on their website - a hub of popular AI-powered applications and open-source models available via a simple API call that you can play around with same-day.
Chronicles of the circuit circus
Apple Is an AI Company Now - by Caroline Mimbs Nyce for The Atlantic. The big pull quote:
“Some of the differences between Apple’s approach to AI and that of the other tech companies can be explained by their respective business models. The tech giants don’t all make money in the same way. Google and Meta control about half of the digital-ad market, and AI-powered chatbots could become just another way to get us to buy things. Microsoft is less in the ad business, but it hopes that adding chatbot functionality to search could help chip away at Google. Amazon’s enormous cloud-hosting business stands to gain from the adoption of large language models (they have to live somewhere!). Apple is a luxury brand, more deeply in the business of making using your computer and phone enjoyable above all else. “So it isn’t surprising that Apple is approaching AI cautiously, with a product-oriented focus,” Gruber said.”
Schumer to call for an 'all-hands-on-deck' approach to regulating AI - by Scott Wong for NBC News. The big pull quote:
Those alarm bells have caught the attention of top political leaders and policymakers. During his visit to the Bay Area on Tuesday, President Joe Biden huddled with eight AI experts in San Francisco, including Jim Steyer, CEO of Common Sense Media; Sal Khan, the founder and CEO of Khan Academy; and Tristan Harris, executive director and co-founder of the Center for Humane Technology and a former design ethicist at Google.”
Cooperation or competition? China’s security industry sees the US, not AI, as the bigger threat - Dake Kang for AP News. The big pull quote:
“China’s tech firms have approached chatbots like ChatGPT with caution, for example, because of heavy censorship, having AI generate politically sensitive content is a no-go.
But ChatGPT begs the question: Should China rush to embrace AI and possibly fall prey to its pitfalls, or tiptoe cautiously and risk falling behind the United States?
Across the Pacific, American tech executives and policymakers are grappling with the same questions. Waves of U.S. sanctions have targeted Chinese chipmakers and AI companies to restrict Beijing’s access to cutting-edge technology. Politicians worry about China’s growing prominence in the field.”
Thanks for joining me for another Future of Product! If you have any thoughts, or want to discuss any of the topics I touched on here with me directly, don’t hesitate to comment here on Substack. Oh and in case you missed it, click here to listen to my podcast with Rebecca Milazzo, my friend and co-worker over @ PlayerZero.
Can’t wait to see you next week!
Thanks for reading Future of Product! Subscribe for free to receive new posts and support my work.