Some days ago, Google announced antigravity*, basically an agentic interface that is up and ready to solve all you could possibly need. Of course, in order to achieve maximum integration, it is required for the program to have complete control of your computer.

Sounds like a dream, isn’t it? Finally, a robot servant, free and ready to complete all your whims. Not so much for the reported developer who lost all his data when it misinterpreted a cache-deleting query**.

Be mindful that, depending on how this is framed, it’s actually a relatively easy mistake to make—and mind you, this is a developer, someone we know almost certainly has the capacity to undertake this type of task. One can only wonder what luck the average user, with half-baked prompts, will have when starting to toy around with credit card numbers, passwords, and, in general, the increasingly digital part of our lives.

For you see, from a cybersecurity perspective, the weakest vector will always tend to be human failure. It seems complicated to the average user, but this is mostly because these are realities that not many of us ever come into contact with. For comparison, red teamers and pentesters (people simulating cyberattacks for a living) utilize programs and terminals that, laughably enough, have less power than most phones today, on obscure systems that affect you regardless of your understanding.

The problem, as usual with the integration of new technologies, is not really at the end of those implementing it. Enthusiasts tend to have at least partial knowledge of the dangers behind a new tool; the issue is the user.

After all, the chain is always broken at the weakest link, which is almost always the user (or at least an uninformed one).

From Google’s logic, this is not really an issue, as they are arguably more interested in the window opened straight into the main controllers of your PC—something that actually undermines your control of your own electronics to an unprecedented level, at the cost of whatever information you are willing to provide for the training of their new models.

At the same time, there are a series of interesting changes coming into the cybersecurity space, also announced by Google***, with their president of global affairs suggesting the adoption of quantum cryptographic standards starting now, as quantum threats loom on the horizon****.

I have on previous occasions warned against the blanket transfer of control to foreign AI systems on a computer*****; a danger amplified by the growing interest of major actors in collecting vast amounts of information, especially with the advent of far more powerful decryption technologies. Which means I will gather it now and open it later. A date which is now considered a lot sooner.As usual, it seems that basic logic is what has abandoned the common user. Ultimately — and this comes from someone who is an ardent defender of digitalization —if you don’t want it online, never put it on a computer in the first place.

Notes:

1.Google Antigravity

2.Google’s Antigravity AI deleted a developer’s drive and then apologized | TechRadar

3.Google Shifts Post-Quantum Encryption from R&D to Government Policy Mandate | The Meridiem

4.Digital security in the Quantum Era

5. AI Agents and the New Erosion of Privacy – _BlnkTerminal

---

The Pocket AI Guide is out!

📙 Amazon US: https://a.co/d/gCHHDax
📗 In Europe Amazon Germany: https://amzn.eu/d/3cmlIqa
(Available in other stores Amazon stores too in Europe)

Check the free resources in this website!