Home >Technology peripherals >AI >Why I left Google AI and chose blockchain
Artificial intelligence is currently the brightest star in the technology circle. AI tools like ChatGPT are now sophisticated and sophisticated enough to pass MBA and medical licensing exams, and almost 30% of professionals surveyed reported using ChatGPT to draft emails or write code.
At the same time, the industry’s view on blockchain seems a bit pessimistic, but this is understandable. After all, with the depression of the Metaverse, countless people have suffered from the collapse of NFT prices and the implosion of major exchanges. Lost almost my life savings.
I have worked in both fields. I started as a manager in Google's Machine Intelligence group, building a product that was used to diagnose eye diseases. But when I came into contact with and understood blockchain, I chose to leave Google and embrace blockchain technology.
Given the current rapid development momentum of artificial intelligence applications and the huge resistance faced by blockchain, I was asked repeatedly why I made this decision. For all kinds of questions, the answer I give is very simple: AI only empowers individuals, while blockchain can empower the collective.
Without specific applications, technology loses value: a hammer seems useless until it is used to build a house.
(1) The main role of artificial intelligence is to handle routine and even complex tasks that people are prone to making mistakes.
(2) The main function of the blockchain is to enable people to collaborate without considering trust relationships.
While at Google, I was responsible for building a product that could assess a patient’s risk of blindness based on images of the back of their eyeballs. Doctors used related tools to generate a dataset of more than 100,000 images and their associated disease states:
After training this dataset, Artificial intelligence can diagnose patients with the following diseases with 90% accuracy. Although the performance of artificial intelligence is on par with professional doctors, it has two important advantages:
(1) Speed: The model can output diagnostic results in a few seconds, while human doctors require about 10 It takes minutes to complete the diagnosis, which increases the efficiency of diagnosing cases by at least 100 times.
(2) Quality: Because the model is trained on a large enough data set, it can perform better than many inexperienced doctors and will not be affected by fatigue in processing tasks.
The impact of introducing artificial intelligence into this workflow is clear: By offloading time-consuming, repetitive image recognition tasks to artificial intelligence, physicians can focus on what they do best: patient interactions. But through this product, I also realized that artificial intelligence cannot replace humans after all. It can only save our time and energy by handling low-level tasks.
The popular ChatGPT promises to give users more help. While writers and developers using ChatGPT can let computers handle the boring stuff, they still need to rely on their own experience and talent in the space of creative thinking.
Of course, until recently, I still believe very much in the power of artificial intelligence to empower individuals. This is indeed a meaningful period, but after getting in touch with and understanding the blockchain, I found my more Things you want to do.
When I joined Google, I was very clear about the value proposition of artificial intelligence, but I only occasionally practiced blockchain. During the epidemic, I used my spare time to do volunteer work in a museum. At that time, one-third of museums were facing bankruptcy, and I hoped to rely on fundraising strategies to help them tide over the crisis. And that's when I encountered an unexpected solution: blockchain.
When people donate to any cause, the donor has no way of knowing whether the money will be distributed as promised. It also has no say in how the funds are allocated, as that is entirely decided by the recipient organization.
The blockchain subverts this model. Blockchain is like any other database, except that no single entity has control over it (decentralization).
Due to its decentralized nature, blockchain ensures that all contributions are transparent. Take fundraising, for example. The traditional model is to track transactions on a central database controlled by a bank or fundraiser. These databases are owner-managed and only have access to them, meaning donors don’t have any visibility. In contrast, blockchain allows anyone to view records on a public blockchain, holding fundraisers accountable for the proper distribution of funds they raise.
Pushing funding entities towards transparency already represents a significant improvement on the status quo. But the true power of blockchain will be revealed when the central fundraising entity is completely eliminated.
For example, when a certain place is suffering from a natural disaster, thousands of people need help from the outside world. But everyone involved in fundraising wants their money to be spent on worthwhile and trustworthy projects.
Under the current circumstances, people’s only option is to donate money to fundraising organizations and then pray that they will use the money as promised. This requires the donor’s unconditional trust in a group of strangers they have never met. people.
But that changes if you can donate to a platform that is managed entirely by donors rather than third parties, which is the core idea of blockchain. In addition to providing a transparent and secure database, many blockchains support an additional feature: voting rights. Logic is built into the blockchain to automatically allocate funds based on voting results.
For example, a blockchain-initiated fundraising campaign might use voting results to evaluate how to distribute raised funds. A donor offers to pay Company X to rebuild the home. Another donor proposed hiring Company Y. If a majority votes in favor of Company Y, funds tracked on the blockchain will automatically be paid to Company Y. In this way, funds are transferred from donors to operators without any intermediaries, and the donor's input is automatically accounted for. As a contributor, in the blockchain you no longer need to rely on any entity to ensure that your output is not misused.
In short, both artificial intelligence and blockchain have their desirable and disgusting sides. These areas require a lot of correct values and user experience to guide them in order to have a chance of gaining positive influence and being adopted on a large scale.
While I still have doubts about most blockchain use cases and products today, I also still believe in the potential of the technology: blockchain allows collectives to conduct themselves without relying on third parties. Production and management. The group's database tracking activity is decentralized, and because voting results are automatically enforced, the collective is able to completely self-organize. At that time, those "leaders" and "managers" who hold all their power will no longer be meaningful, and they will no longer be able to extract resources from the collective they control.
We have no way to predict whether blockchain and artificial intelligence will have greater social impact in the future. They are just tools, and their output depends entirely on how the user uses (or abuses) the tools.
Original link: https://hackernoon.com/why-i-left-googles-ai-division-for-the-world-of-blockchain
The above is the detailed content of Why I left Google AI and chose blockchain. For more information, please follow other related articles on the PHP Chinese website!