Rewritten in November 2026. We first published this article in May 2020 as “5 Predictions from the Past About the Future of Learning have Already Happened”.
An honest word about predictions
Everything below is opinion. These are reasoned expectations, not facts, and they come with no statistics. People who predict the future of learning are often wrong in detail. The old version of this article claimed that earlier predictions had already come true. We cannot check that for you, and some of the changes it described are obvious without help: more learning happens online, and more of it happens on a phone.
So we use a different test. We ask what is likely to stay true, and what a learner can do now, whatever happens.
Expectation one: learning moves to where you already are
It seems likely that people will keep learning in small pieces, in the gaps of a working day, on whatever device is at hand. A short exercise on a train is easy to start. What this means in practice: design your week so a ten-minute session is always possible, and keep a longer session for the harder work.
Expectation two: tools get better at feedback, but people still decide
Software already corrects some kinds of errors and will probably correct more. It is reasonable to expect more personal practice and faster feedback. What we doubt is that it will replace the person who tells you what a sentence means in context, or how it lands with a particular listener. Use the tools for repetition, and keep human conversation for judgment. See our note on choosing the right tools for how to decide.
Expectation three: access grows, attention stays scarce
More material will be available to more people, in more languages. The scarce thing is not material but attention and a plan. A person with a thousand free lessons and no routine will learn less than a person with one lesson and a schedule. Choose less, and finish it. The advice in staying on the path of learning covers this.
Expectation four: learning becomes more personal
It seems likely that courses will fit more closely to the learner’s own work and interests rather than to a general syllabus. This matches an old idea: people learn faster when they care about the content. There is a related, much-debated idea that people have different strengths in learning, which we discuss in our piece on multiple intelligences. Treat that as a prompt for experiment, not as a rule.
Expectation five: speaking in your own voice matters more
As machine translation improves, it may seem that learning a language is less necessary. We expect the opposite for people who work with others. Translation can carry information. It is much less able to carry trust, humour and a personal style, which come from speaking for yourself. This is a belief, and you can test it: compare how you feel when a partner addresses you in your language, even badly.
What if you want to prepare
Do three things, whatever the future brings. Set a clear goal for what you want to say and to whom. Practise speaking aloud, with people, every week. And review once a month what is working. A learner who does these is ready for most changes in tools. If you want structured practice with a coach, see the academy. When your motivation drops, what keeps an adult practising is a good place to look.
Read each expectation as a question to test against your own experience. If one of them does not fit your life or your work, drop it and keep the others. Tell us where we are wrong in a year or two. We would rather be corrected than be certain.