Binggao is studying how to build a web page.
This is a bit difficult for a two-year-old ragdoll cat. Please wait.
If you are impatient, you can contact Binggao at [email protected].
Remember to bring some freeze-dried salmon!!!
Here are some articles written by Binggao before, you can take a look first:
Futurist Ray Kurzweil pointed out the “Law of Accelerating Returns” in his 2005 book The Singularity Is Near: technological progress grows exponentially, implying that artificial general intelligence may arrive very soon. Nobel Prize in Physics laureate Geoffrey Hinton is deeply worried about the future, warning that super-human, unaligned AI could spiral out of control—or even replace humanity—bringing unpredictable shocks and consequences for civilization. Confronted with this, Nick Bostrom argues that we should proceed cautiously with AI and, in parallel, augment humanity through technology.
There are many ways to achieve human enhancement. Exciting, almost sci-fi methods include gene editing and smart prosthetics. Yet these approaches still feel distant, while the challenges posed by AI are already at our doorstep. A more readily available path is to use AI itself to enhance human beings.
To augment an individual via AI, one must possess data about that person. Within AI research, the three pillars are generally considered algorithms, computing power, and data. In the human-enhancement scenario, the data requirement is special because the data must pertain to the specific individual. For example, if we want an AI to handle routine messages to boost someone’s work efficiency, we first need to supply examples of how that person has handled similar messages so the AI can learn their style, habits, and methods.
The route of collecting and leveraging data to achieve greater intelligence has been validated repeatedly. Industry has pursued the “informatization → digitization → intelligentization” pathway for years, and related theories such as Management Information Systems are well established. Traditionally these frameworks are applied at the organizational level; the new task is to migrate them to the personal level. Today, an **AI agent—perhaps even multiple agents—**can help process information and carry out simple actions, enabling people to complete tasks with higher quality or greater speed. In this collaboration, data and information are the medium.
Ideally, we would collect the most comprehensive, time-series data possible about the individual. This includes internal physiological data—body temperature, heart rate, blood-oxygen saturation, glucose levels, EEG, EMG, hormone concentrations—and externally observable data such as speech, movements, behaviors, and interactions with others and objects.
However, current technology limits the scope of what can be collected. Many imagine brain–computer interfaces (BCIs), yet invasive BCIs have only a few experimental cases so far, with unclear side-effects; non-invasive BCIs lack the precision to meet practical needs. Some data could theoretically be captured, but no products exist, or the devices are too bulky for daily life. Consequently, only a small subset of desired data can actually be gathered continuously.
For individuals, data collection is urgent: humans grow, change, age, and eventually die. Much personal data is non-renewable—miss it once, and it is gone forever. Our biological memory pales against silicon storage; we retain only a fraction of what we experience each day, and forgetting, blurring, or mixing memories is routine.
In fact, well-collected data can be used long afterward, and its value may rise over time, not fall. Even if current technology cannot fully exploit the data, any new technique can immediately make use of it. Therefore, one should start collecting whatever data one can right now.
Against this backdrop—humanity needing enhancement to face the future, and personal data being essential—I embarked on what I call an “external memory” project: collect every piece of data about myself that I possibly can, store it, and eventually have external systems process it to help me optimize decisions, or even handle tasks directly, thereby boosting my own intelligence. Although my current focus is data collection, I also experiment with processing and utilizing it, aiming to consume not raw data but the information and knowledge extracted from it.