Paolo Malinverno is SVP of Strategy and Innovation at Salesforce. This article by Paolo is based on his talk at an Apidays event in Paris.
Five years ago, many people were discussing trends that would be set over the next five or ten years. But, the analyst within me was rebelling: those things were happening already, they were not trends for the future, they were present trends at that time. So, I thought of stepping back and analysing the megatrends that will cause a lot of API consumption.
I think the one true thing for the future is uncertainty, which we must cope with. The future is a recent invention. If you think about it, a hundred and fifty years ago, nobody thought about the future because there were too many problems in the present to deal with.
Easy Truths for the Next 5 Years in APIs
I have selected a few things that will come in the future and will discuss the API consequences for those.
- Everybody takes APIs for granted. Around twelve to thirteen years ago, everyone talked about e-commerce and the ability to buy over the Internet. Now, everybody buys over the Internet, and nobody talks about it.
- We consume APIs more than we design APIs. This in itself is a kind of reuse. People will want to use APIs for free, making monetizing them difficult.
- The absence of standards will continue. When there are too many standards, there is no real standard.
Megatrends
- Organizations will create APIs that business technologists will use. The APIs will be designed, coded, and tested by a group of people. Consider your API to be a building block, like a microservice or a piece of user interface. But the people using them will have no notion of the load of an API executing on a system. They will just look at the API as a black box, that does the job they need done.
- Business technologists are non-IT people who know technology and have business. These people vastly outnumber IT people and will build many useful automations, which frequently can be turned into APIs. Business technologists will have good ideas and may advance your digital strategy, but they need governance. You do not want to put them down, but you also want to get the right ideas in the automations to make them effective and clean.
- If you include external APIs in your developers’ portal, rate-limit them if you pay for them on a consumption basis.
Sustainability
Sustainability is a big thing. Every company will have to report and manage greenhouse gas emissions. According to the Paris Agreement, which was done here about seven years ago, each nation will have to respect these limits on greenhouse gas emissions. So, data on greenhouse gas emissions will have to be calculated, which is difficult. Over and over. This data changes if you move a data center. If you move it, say, from Ireland to Norway, your greenhouse emissions will go way down. The consequence on APIs is that you will have a lot of massively data-intensive APIs that will be called to get accurate emissions data out of things that might not even have an API; it might be a piece of machinery and maybe something else. Expect this to be big and to become bigger and bigger as the sustainability targets tighten.
Many companies will not be able to be as sustainable as they want as quickly as they want. So, they will buy sustainability credits. This will be done through APIs too.
AI Job Augmentation
Everybody is scared of AI getting rid of jobs, but AI also creates jobs. In the coming years, there will be a lot of AI job augmentation. Humans will use AI to do the job more effectively, and this is most likely to happen.
But there is a danger. How can you trust the AI algorithms that will be packaged in an API? For example, you have 500 dollars to invest. You have an API looking at everything and ensuring that you will have your best return on a six-month basis. Can you trust it?
If you use AI APIs, assess the risks. Ask yourself five key questions –
- Is this real AI? There is a lot of fake AI out there
- Who financed the development of the API?
- Who trained the algorithm? Can you add to that training? Does the algorithm actually learn?
- Can you trust the results of the APIs?
- What are the downstream consequences and risks of skewed results?
The problem is that AI is not ethical. It doesn’t have a notion of ethics or right and wrong. We need to teach AI to be ethical. This is the fundamental point.
The Metaverse
As quoted in a Salesforce ad, “Space is the boundary of human achievement, the new frontier. It’s not time to escape. It’s time to engage. Time to plant more trees. Time to build more trust. Time to make more space for all. While the others look to the Metaverse and Mars, let’s stay here and restore ours. Yeah, it’s time to blaze our trail.”
When I talk about Metaverse, I make two fundamental warnings. First, I would rather fix this one before creating a new world. Because if we create another one, it will have the same problems or even worse. The second thing I would say is I never trusted social networks. I never trusted the billionaires on top of social networks. Ask yourself if, in 10 years, you will have the same social networks as you have today because you won’t.
Web3, Immersive Technologies, Metaverse
The Internet was a success because it was a net of nets. Web3 and Metaverse are complementary concepts evolving in parallel. Similarly, immersive technologies will become popular for training, remote expert guidance, remote inspection, and remote collaboration.
With information stored in multiple locations simultaneously and no single authority or ownership, Web3 will be either API hell or API heaven. To me, Web3 will happen quicker than the Metaverse. The Metaverse will continue to evolve. Oculus Immersive technologies will also evolve quite a lot. There will be a huge investment in metaverses. What worries me most is, what if people prefer living and working in the Metaverse that in reality? That will cause a lot of alienation issues. As metaverses form, expect gazillions of APIs being created and used.
Digital Twins / Digital Humans
Digital twins are digital copies of other people or pieces of machinery or even organisations. They have been around for years. Digital humans are AI-powered human-like virtual beings who can connect to any digital entity, take decisions and interact with humans using verbal and non-verbal cues.
Today, everybody uses the APIs into digital twins, either to test a specific thing or call other APIs into the real twin to change things. Digital twin APIs are used to test “what if” scenarios, provide advanced customer care, plan predictive maintenance, etc. Physical twin APIs are called to perform real-life actions. Digital humans will run algorithms to simulate human behavior and will be active, and call APIs too. This will give rise to a digital humans’ economy.
Conclusion
To conclude, the real value is Ethics. Apply ethics; it is fundamental, especially for AI, especially for the algorithms hidden in the APIs you will design or use. Computers are not ethical; we need to teach them to be ethical and upfront, and honest when we program them.
