Why New Technologies Aren't the Answer to All Our Challenges
Written on
Chapter 1: The Allure of Technology
Throughout history, tales have circulated about market predictions, such as the one involving John F. Kennedy’s father, who allegedly foresaw a stock market crash in the 1930s based on tips from his shoeshiner. Whether true or not, this narrative embodies a common theme in discussions about market downturns.
As a product manager and software engineer, I may not trade stocks, but I observe similar patterns in technology. Often, individuals approach me with grand ideas for apps, already branding them with names like “Mercury” or “Borealis.” Despite being essentially a database and a user interface, these concepts inspire enthusiasm. However, the conversation typically shifts when I inquire about their current computer inventory, leading to an eager query: “Could we incorporate AI for that?”
The optimistic tone is somewhat amusing, as if I’ve been hoarding AI capabilities. The expectation seems to be that I can magically deploy AI to handle this task, as if I’ve been intentionally withholding it.
The crux of the issue lies in the term “artificial intelligence.” Influenced by science fiction, many presume that AI has sentience, especially with warnings from figures like Stephen Hawking and Elon Musk about rogue AI posing threats. In reality, today’s AI primarily excels at recognizing patterns in data.
AI is predominantly about matching patterns. We provide it with data and specify our goals, but it lacks true intelligence. For instance, an AI's potential threat level is akin to that of a harmless game like Minesweeper, which becomes dangerous only under specific, unusual circumstances.
To illustrate, scientists at Stanford tried to train AI to diagnose skin cancer. While initially successful, the AI’s accuracy dwindled because it learned to associate the presence of a ruler in training images with cancer detection. This illustrates the adage: garbage in, garbage out.
If AI isn't the panacea for our problems, what is? Perhaps the Cloud, virtualization, or blockchain?
Section 1.1: The Reality of Chatbots
Years ago, I attended a tech event hosted by a prominent company known for its red logo. At the time, chatbots were all the rage, seen as potential substitutes for customer support.
“Can we implement a bot for HR inquiries?” colleagues would ask.
“What tasks should the bot perform after receiving a query?” I would respond.
The disappointment on their faces was palpable. It was as if I were crushing their hopes by clarifying that no universal technology could answer specific HR questions without structured processes.
During the event, the company showcased their chatbot capabilities, demonstrating how to integrate with platforms like Facebook Messenger to provide predetermined responses. However, when we explored the coding behind it, it became apparent that the logic was merely a series of conditional statements. This revelation diminished the allure of these bots, ensuring that support roles remained secure.
Section 1.2: The Voice Assistant Dilemma
With the advent of virtual assistants like Alexa and Siri, the desire for voice-activated solutions has surged. Yet, is it truly more efficient to verbally list expenses rather than typing them?
I believe the fascination stems from portrayals in media, such as Star Trek, where Captain Picard orders a “Tea. Earl-grey. Hot.” This brings up questions about the default temperature. In reality, even in the 1980s, a simple button could have sufficed for drink orders.
The critical takeaway is that technology should be applied to the right problems. Just because machine guns outperform knives in warfare doesn’t mean they are suitable for spreading butter.
Chapter 2: The Limits of Technology
I’m not a traditional Wall Street investor, but I recognize when technology gains enough recognition that its terminology loses impact. When terms like “magic quadrant” circulate widely, I often receive inquiries from non-technical colleagues.
There was a time when the internet was the must-have. Then, apps became the next big thing. Each new development was supposed to replace the previous solution.
However, the stark reality is that implementing advanced technology requires substantial groundwork. There are no shortcuts. Regardless of how innovative a technology appears, success relies on diligent efforts behind the scenes. Even the most sophisticated AI requires extensive data cleaning and preparation, which isn't nearly as exciting to discuss.
The first video, "Why We Can't Fix Bias with More AI," featuring Patrick Lin, delves into the complexities and limitations inherent in AI systems, highlighting that simply adding more AI won't address underlying biases and issues.
The second video, "Technology Won't Solve All Your Problems," explores why relying solely on technology for solutions can lead to oversights, emphasizing the need for human involvement and critical thinking.