The idea of “AI liberating humanity” is such a seductive illusion! > 4th Part

Oh, wow. The first one. I don’t like to argue, but when I show Google Lens the plants on my terrace, it almost always shows me a different name for the same plant! If you know what I mean… I can never completely trust it. Second, I don’t understand how AI can analyse soil. I don’t have a lab, so how does it do it?

This happens because Google Lens is not a true agricultural diagnostic tool. It is a general‑purpose image recogniser trained on millions of random photos from the internet. When you show it a plant, it tries to guess the closest match from its huge database, even if the match is weak. That is why the same plant can be identified as three different species on three different days. It is not designed for precision; it is designed for “best guess.”

Real agricultural AI works differently. It is trained specifically on plant diseases, leaf patterns, and crop types, using controlled datasets from agronomists. Instead of guessing, it compares your photo to thousands of verified examples. That is why specialised plant‑diagnosis AI is far more reliable than Google Lens. The problem is that these specialised tools are not widely available or well known, and many are locked behind expensive subscriptions. So the technology exists, but it is not reaching the people who need it most.

AI does not “analyse” the soil in the chemical sense. It does something more subtle and surprisingly useful. AI can estimate soil conditions by combining several types of information:

First, it uses satellite data. Modern satellites can detect moisture levels, temperature, and even organic matter on the surface. AI models interpret these signals and translate them into practical advice for farmers.

Second, it uses simple sensors. A small, inexpensive stick‑shaped sensor placed in the ground can measure moisture, pH, and temperature. The AI interprets the sensor readings and predicts how the soil will behave in the coming days.

Third, it uses pattern recognition. By looking at the color and texture of the soil in a photo, AI can estimate whether it is compacted, dry, nutrient‑poor, or suffering from salt accumulation. It is not perfect, but it is surprisingly accurate for everyday decisions.

Fourth, it uses climate history. AI can combine years of local weather data with your current conditions to predict how the soil will respond to rain, heat, or drought.

None of this replaces a laboratory test, but it gives farmers continuous guidance without needing expensive equipment. It is like having a digital assistant that watches the land and warns you before problems appear.

“Real agricultural AI works differently.’’ = how much can cost? on what criteria would be a price approximated ? Whoa for the second, i think one need ‘’millions for such an app! If ever would be possible only through an app! Maybe i am wrong!

The price varies enormously, and it depends on what kind of AI we are talking about. There are three main categories.

The first category is simple AI apps for plant disease detection. These can cost almost nothing, sometimes just a few euros per month, because they run on your phone and use cloud models. They are not perfect, but they are far more accurate than Google Lens because they are trained specifically on agricultural images. The price is low because the company spreads the cost across many users.

The second category is professional agricultural AI systems used by medium or large farms. These can cost hundreds or thousands of euros per year. They include satellite monitoring, weather prediction, soil moisture analysis, and crop‑growth modelling. The price is based on the size of the land, the number of features, and the frequency of updates. These systems are powerful, but they are too expensive for small farmers unless subsidised. The third category is AI integrated with physical equipment, such as drones, sensors, or automated irrigation systems.

These can cost several thousand euros because you are paying for hardware plus the AI software. This is where the price becomes unrealistic for small landowners. So the cost depends on the complexity. The simplest tools are cheap. The advanced ones are expensive because they require servers, data, maintenance, and sometimes hardware.

It feels like it should cost millions, but the trick is that the expensive part — the satellites, the climate models, the global data — already exists. Farmers do not pay for satellites. They do not pay for climate supercomputers. They only pay for the app that interprets the data. The app itself is not expensive to build once the infrastructure exists. The AI model uses information that is already collected by space agencies, weather services, and environmental sensors. The app simply translates that information into advice for the farmer.

This is why it is possible. The cost is not in the app; the cost is in the global systems that governments and large institutions already maintain. The farmer only pays for the interpretation layer, which is relatively cheap. The technology exists. The problem is that it is not reaching the people who need it most. Small farmers often cannot afford subscriptions, hardware, or stable internet. This is why AI has not yet transformed agriculture in the way it could. The potential is real, but the distribution is unfair.

Let me give you some practical examples. Plantix is one of the most widely used agricultural AI apps. A farmer takes a photo of a leaf, and the app identifies diseases, pests, or nutrient problems. It is far more consistent than Google Lens because it is trained only on agricultural images. The basic version is free, and the more advanced features cost only a few euros per month. This makes it accessible even for small farmers. It is not perfect, but it is reliable enough to prevent crop loss and reduce unnecessary chemical use.

Another practical use is planning. AI can look at the local climate, the type of crop, and the expected weather for the next weeks and suggest when to sow seeds or when to expect harvest. This is especially useful for small farmers who depend on timing to avoid losing their crops to heat, rain, or pests. It does not require sensors or machines, only a phone with an app that interprets the data. These tools do not replace human knowledge. They simply add a layer of guidance that helps people avoid mistakes and protect their plants. They are not perfect, but they are real, and they can make life easier for anyone working with the land, whether it is a large field or a small terrace garden.

Farmonaut uses satellite images to estimate soil moisture, crop stress, and vegetation health. The farmer does not need sensors or machines. The app simply interprets satellite data and gives advice about watering, fertilising, or protecting crops. Prices start at around ten to twenty euros per month for small plots. It becomes more expensive for large farms, but for a small landowner it is surprisingly affordable. The reason it works is that the satellites already exist; the app only interprets the data.

You mentioned that such technology must cost millions. The truth is that the expensive part — satellites, climate models, global datasets — is already paid for by governments and space agencies. The app only interprets the data, which is why the price can stay low. The real challenge is not the cost of the app but the fact that many small farmers do not know these tools exist or do not have stable internet access.

Apps like Plantix or Leaf Doctor are more reliable than Google Lens because they focus only on plant health. But even they are not perfect. The best way to use them is as a helper, not a judge. Take a photo of a leaf, read the diagnosis, and then compare it with what you see in real life. If the app says “fungus,” look for spots or powder. If it says “nutrient deficiency,” check the soil or watering habits. It becomes a conversation between you and the plant, with the AI giving hints rather than orders.

Many apps can analyse a photo of your terrace and estimate how much light each corner receives. This helps you place plants correctly: sun‑lovers in bright spots, shade‑lovers in protected areas. It is a small thing, but it makes a huge difference in plant health. And many others my friend!

Well, I must say. I am overwhelmed! Good to know though! Thank you. See you next time my friend! And don’t dive to much! Starting from a polemical position of the richest people on this planet and getting to all of this! What a conversation!

Aha ha ha ha! Sleep well!

©Ca De Luce> MINDFUL MIND Medium Blog 2025. Unauthorized use of text or media is not allowed. All images and photo are fulfilling the copyrights regulations. Much obliged to you all!