Next industrial revolution? May -go. A simple test of 1950 tells us soon enough. And if yes – what do all unemployed doctors, lawyers and programmers do? Here is one suggestion.
“Siri, do you take my job?”
“Sorry, I don't want to answer this question. Can I do anything else today?”
Technology is essentially an increase in efficiency. What was done with more time and resources – can now be done less. Do you have to make a copy of the book? Months for the scribe, seconds in the printing house. Do you have to go from Rome to New York? Weeks by boat, for hours. If the change is large enough – width and depth (the number of people affected and the number of resources saved) – can be considered as a revolution *.
* It also works differently. If the technological healing is not large enough – the one who offers it fails. People do not change habits for a small healing. Of course, the disks are small, do not skip and can register compared to CD-ROMs, but… I wait for the iPod.
Time “liberated” also means that people paid for that time (their work and their ability to acquire their abilities) is now “free”. Which is a nice way to say “out of work”. On the sunny side – new jobs are also created. If the technology is crucial enough – brand new ecosystems. Think about a trip, computers, internet.
If the change is gradual, the professions die at the pace that their passengers retire or manage a career shift. If the change is too rapid – can cause social instability (between people with capital and people based on their labor). The dark times of history are attributed to such changes – so we hope we do not stand in one.
Now that I have your attention and your night sleep, we will discuss AI. How do we find out if this is a “revolution” or just “reve”? What new new professions/ecosystems will appear? Are big money earned in brave new ways?
How to find out if AI is really a revolution
We have a short memory and have never listened to history.
Otherwise, it is difficult to explain the shouts: “Stop AGI before it's too late!” And “Stock Nvidia chips under your mattress!” If we were to actually murmur “AI again?” Really? “. AI -D Rabelet as a change of world in the 1950s. And then the 1960s, the 70s, back in the 1980s, reappeared in the 90s, then in the 2000s and finally. Of course -” This time it is different “due to new algorithms, data and computer power. But is it really? We do not see mobs for doctors, lawyers, teachers and programmers with burning torches, markers, stethoscopes and keyboards.
So how do we find out? Fortunately – is the test since 1950. Turing Test. If the current AI takes over it – will probably be followed by a major change. The Turing test requires that a person who is hidden in a “thing” does not make the difference between the answering machine and the person. If this happens – according to the test – the machine is considered intelligent. I would take this step forward – if any The person has a conversation “with a thing” and cannot tell if the other party is a aware Man or machine – we have arrived. Why “some”, “conversation” and “aware”? Because of the expertTo do, finishing and truth – There is no real value.
Are we there? Not yet. And a new AI winter can arrive before it reaches there. The amount of data, computer power, algorithm – may not be ready. AI is a nice feature-like looking for automatic, personalized feed and advertising algorithms. They are useful, they make a lot of money – but they are not a social revolution. Remember “Deep blue”.
The apocalypse of AGI and the paper valve is really a question. Algorithms are not even targeted there (we don't know where to target). The Turing test is in the hope of the hopeful intelligence a bit (pun intended) naive. Maybe if we sprinkle a little dopamine on GPU …
But let's imagine that the Turing test will be completed in a year. For a doctor to have a thorough medical debate with the “thing” and do not know whether it is a machine or a peer. The fact that the machine actually makes less mistakes than a person who gives a diagnosis and treatment recommendation. It can teach you Japanese while commuting. You can describe what the software should do and it works on your device.
At that moment – learning for years, working and working for hundreds of dollars per person and understanding for their professions. But it is only half of the Turing equation.
Questions with questions – the need for diagnosis, legal advice, software idea, a desire to study or travel abroad – still exist.
These are all different tasks. And these tasks are performed with different tastes for different people, at different times. The Australian surfer in his 20s is not the same as retired by the Japanese engineer – even if they both ask for travel recommendations.
Enter your companion design.
What is the design of a companion?
“Brahhhh! Surf is on South Beach! Rolls 60 to one! Would you like me to send gonzo and frenchie to see if they are down?”
“True, White's collection is more dada than Blaue Reiter, but I think you have it interesting. Do you want to give you a quick overview of the works and why are they intriguing?”
“Nan! You want me to teach you Japanese by focusing on Manga? You need a lot of dokonjou, but ok. Ganbatte!”
When AI passes a Turing test – we want to interact with people. And because we (different people) communicate with different people in different ways – so is our desire to be. We want our travel companion to respond to our preference to extreme winter sports and be behind guides, gears, insurance and weather. We would prefer it to know what the difference is between the frame, the plate, the railway and the volumes without the need to “specify”. And in the context of work – we assume that it knows what Dau, CTR and CLTV are and use them in the right context.
In other words, we want our AI companions to be “culturally customized”. And someone has to create those companions. AI can't guess who you are and how you prefer to communicate in every context.
Skills needed to design your companions
Co -design is a mixture of product management, product design, psychology, sociology, cultural studies and art. This is the profession of building a machine.
Fine-tuned companions can use users through their tasks, which is effective and enjoyable. This is where all of the above abilities come into play. An interesting twist in the history of human education – the humanities come back.
Because there are many subcultures and each person belongs to several times at any given time (work at work can have very specific music and fashion flavor) and move between them over time (choosing new and developing hobbies, changing flavors, trying new things, changing marital status, career), many characters have to be done.
First – comes sourceTo. Big data sets. Some goals (such as internet) and some subjective (such as our conversations, e -mails, photos, interests in social media we go and see).
Then – companions must be able to connect with these pools – and they do AI GatewaysTo. These gates – today are names such as Chatgpt, Gemini, Claude, Meta Ai, Siri (joking. But probably something new in June) is what has understood unstructured data to be fed.
Then – comes to a companion or Tools for creating signs To. They allow the character of the companion to be shaped. Language, vocabulary, slang, attitude and knowledge that this help attracts.
Finally – companions must be implemented In the wild.
- You can Part of the existing tools – Fashion, travel and politics on Instagram and Twitter.
- Is adapted to a specific business Needs-Lunch-African Factory Production Floor Toyota Engineer.
- Drives on top of Lens with a mixed subjective Sources-Tei Family Digital Doctor, Your Marathon Coach, Geography Teacher, Stylist, Dating Counselor, Planting Optimization Guide.
- Going to be alone Companions – someone to talk to, without a judgment or demands.
The use of these tools (assisting the technical tasks of the employees AIS for interview) is co -designers. They understand the need for human beings, build a companion character to respond to it and connect it to the sources they require to succeed.
Show me money!
That a companion agent “pure” in a commercial van (unlike paid search results) – it cannot be interested in the parties to their conversation with the user. But that does not mean that it cannot make money. To achieve and achieve a recommendation, you need to “close” by money-changing (where to go on vacation, what flight to book, what clothes to buy, what kind of apartment to rent).
It may seem a bit of a time trip, but in fact, a similar company. Amazon. Under certain conditions, Amazon, which is not neutral, does not care about the purchase of visited visitors (thus the director may be completely neutral). It earns money from everything so it pays to the Commission regardless of. For Amazon, it's simple – they are all stores, but the world is still a bit bigger than their warehouses. So how do my companions make money?
They register as an ad-hop subsidiary. Each time they refer to the company – they negotiate through a processing company and assign them the client referred to. Without pre -closed the transaction, or if the companion works for “preferred merchants”.
Think about it as a hotel doer. Those who really care about their customers' preferences. Then give the perfect personalized recommendation. They invite you to book and just before hanging-they ask you to put their name next to the client's name. “Mr Concierge referred to”.
They will come to collect later.
It's a winding. Turing, companions, money
If AI will soon pass a professional Turing test, we must be able to give it a character. The appropriate character for customers in terms of these diverse needs. It is shaped by humanities specialists, members of the new handicrafts that understand people, needs and culture. It gives an amazing value while earning a lot of money to its creators.
The future is not that dark.
But first we have a small test to pass. If we fail, we are back in the refrigerator – right next to the chain chain jar.
