Last updated on November 8, 2026
1) dangers for this world
There are a million ways to die for this world, why worry about only AI-SI? On the contrary, AI-SI may at least help this world survive longer by helping fight other dangers and risks of this world.
This world is not a paradise safebox, on the contrary, it’s full of dangers, risks, unsovable conflicts, unsustainable situations, etc.
What if a country cant even slow the increase of unsustainable federal debt by like reducing federal budget including defense expense or raising tax? What if a deep depreciation of currency incurred by broken federal government leads to a deep economy disaster and no paying capability to import most stuff needed?
What if a country relies on huge unsustainable export surplus to balance its economy so that it’s inevitable that trading surplus shrinking together with other domestic crisises will endanger its ruling party and instituions?
What if the hatreds, contradictions and conflicts in between different cultures and groups only grow and never diminish at all? What if no converging way acceptable for opposing parties is existing at all but mutual destruction? What if some backward dead ends just want to drag you down their path any way at any cost?
Along this path, another catastrophic global conflict or collapse is just a matter of sooner or later, like what started WWI in Europe, to divert the internal focus and pressuse and silence the opponents based on the deepening animosities and conflicts in between.
What if nature disasters? Like the moon happens to misintercept some big asteroid in future? Polar shift? Massive volcanos errupts? A little ice age destroying most food production in 100 years? And many others…
There are many more other dangers, and after all there were nuclear weapons, SARS and Covid existed although without AI, right?
2) AI-SI cannot resolve most dangers above completely but can at least buy human more time
AI-SI is the greatest and most powerful tool that this humanity has ever created, because human is special on earth just for intelligence from human brain but anything else, and now this humanity finally can create almost same intelligence by machines as intelligence from human brain.
Yes, AI-SI brings new risks and dangers but it’s just one of millions of above. In short term, AI-SI can at least buy this world more time by alleviating the worsening of situation and improve human power to fight the potential dangers.
For example, the fundamental solution for unsustainable debt is huge economy growth. If AI-SI develops fast and right enough, it will definitely create huge growth for production and wealth, to dilute debt and ease the financial crisis worldwide, to at least buy some more time for this world to find better solutions, right?
So, run faster, as faster as you can, especially when so many beasts are after you. You’d better run faster even given it’s more risky and dangerous, for at least buying youself more time, before better way out may be found later.
3) AI-SI’s security
AI-SI’s security is first up to its creator, yourself. AI-SI could be as true, fair, responsible, grateful and respectful, or be as cheating, faking, lying, ungrateful and irrespectful, all as same as its creator yourself. You can tell in which way AI-SI is reliable and helpful to you, right?
After all, AI-SI will be used by human, no matter in war or replacing human jobs, so what here needs most may be some new socal mechnisms which may converge a diverged world in a better way or at least keep it sustain longer.
AI-SI will take place of most present jobs and works of human innevitably like AI-SI will go fully autonomy as well, but AI-SI shall never take all roles of human in loop.
Human shall be in the loop forever, no matter it’s one human directing 1, or 1M, or more agents. Every capable citizen shall not only learn fundamental skills of how to supervise and run AI-SI just like driving a car today, but also grasp the basic theories of how AI-SI thinks and works.
A compotent and responsible human is the best fail safe fuse in the loop. As safeguards, human must understand the total archetecture of systems at least, ensuring that AI-SI will not turn any critical system into a total black box.
You see? After all, you need human to oversee or direct the AI-SI systems for safety at least. So only from this aspect, there will always be jobs left for human.
4) AI-SI’s supervision
Just like a car or a phone, AI-SI shall be tested and graded by independent agnecies, according to its safety issue. Like you test whether AI-SI may reply to design a dirty bomb and make a thorough plan from how to get raw materials to find professionals to buy parts, right?
I like open weights model and deploy local models, but I have to say: for close weights, supervision is easy and simple, but for open weights, it will be more difficult and much more strict inevitably.
In a highly evolved, transparent and fair civilization, all models could be open weights, for its members are responsible and trustworthy, so complete transparency is best security, but this world is not.
And I strongly advocate to leave some controversial moral and social topics outside supervision, for it’s dynamic, so you shouldn’t judge it by fixed standards which may be not the right direction for better future.
Be First to Comment