ZeonArka

Former Quantitative Analyst @ Galvani Capital LLC (2019–2025)

[ZeonArka @ZeonArkaym]

The Real Question Isn’t Just Why Jensen Huang Left
The discussion around Jensen Huang’s departure can easily become a story about personalities, corporate drama, and stock prices.

Those are legitimate news topics. But there is another question—less dramatic, perhaps, yet potentially more important to everyone:

As we become more dependent on AI, who decides how we are allowed to use it? And if we disagree, will we have somewhere else to go?

You do not need a technical background to understand this question. Nor do you need to decide in advance which companies are good or bad. Start with a simple thought experiment.

Imagine Someone Could Replace Your Teacher Overnight
Suppose you find an AI tutor that works exceptionally well for you.

It understands your learning habits, explains concepts you struggle with, and helps organize years of notes. Over time, it becomes an important part of how you learn.

Then, one day, the company providing it announces:

The old version is being discontinued. The new one may be more capable, but it explains things differently. Some features you relied on are gone. Others now require a more expensive subscription.

You do not necessarily think the new version is bad. It simply no longer works as well for you.

But you cannot keep the old tutor. You cannot ask someone else to modify it for you. Even if you can export your conversations, that does not mean you can take the system you worked with along with you.

Only then do you realize: what felt like a tool you possessed was actually a service someone else could change.

This is a hypothetical example. But it helps explain why having a workable alternative matters.

What Does Open AI Have to Do With Ordinary People?
For many people, “open source” sounds like something programmers argue about.

But from a public-interest perspective, the most important promise of open models is not that everyone will become a programmer, or that every service will be free forever.

It is that society may retain another way forward.

Imagine a school that does not want to send all its teaching materials to an outside platform. A small business that cannot absorb constantly changing service terms. Or researchers who find that a model does not serve their language or cultural context well.

Can they obtain an alternative they can run independently, examine, and adapt?

If they can, they do not have to place all their hopes in a vendor’s next update.

Of course, being able to download a model does not automatically make it affordable or practical. Hardware, expertise, maintenance, and funding still matter.

The question is not how often a company uses the word “open.” It is whether a usable alternative actually exists.

We Do Not Have to Assume Bad Intentions
At this point, the discussion often takes a familiar turn:

Are you against commercial AI? Do you think every company is untrustworthy?

Neither assumption is necessary.

Companies can create valuable products and reasonably charge for them. Closed models can provide excellent services. Open models can have serious flaws.

The question is different: as a company’s services become more essential, does society have meaningful ways to hold its power accountable?

We do not need to believe landlords are bad people to discuss tenants’ rights. We do not need to believe every banker is dishonest to require financial oversight.

Likewise, we do not need to prove that an AI company wants to harm the public before discussing users’ rights, alternatives, and public accountability.

In fact, the most concerning outcome might require no coordinated plan at all.

A company wants higher profits. An institution wants lower costs. A user wants greater convenience. Each decision may look reasonable on its own.

But if those decisions eventually leave schools, businesses, and public agencies dependent on a handful of systems they cannot realistically replace, what happens when they need to say no?

When AI Shapes More Than Productivity
If AI only helps rewrite an occasional email, a change in service may be a manageable inconvenience.

But imagine it becoming deeply involved in educational feedback, workplace decisions, explanations of current events, and access to public services.

At that point, how it answers, what it prioritizes, and when it refuses are no longer merely questions of product design.

A distinction matters here: not every refusal is censorship, and not every restriction is unreasonable. Protecting privacy, preventing fraud, and limiting dangerous assistance all require boundaries.

The questions are: Who sets those boundaries? Can the reasons be explained? Can mistakes be challenged and corrected? Can users choose a different, reasonable approach?

If only the provider can answer those questions—and everyone else must accept the answer—society may be acquiring a new dependency alongside a useful technology.

Why Are These Questions So Often Missing?
Personal stories are easier to follow than institutional arrangements. Stock prices are easier to turn into headlines than the cost of switching providers.

It is understandable that the most dramatic parts of a story attract the most attention.

Still, we should ask: What does that framing leave out? And who benefits when certain questions go unasked?

If the entire debate becomes a referendum on whether one executive deserves sympathy, we may forget to ask what protections society needs regardless of who becomes CEO.

If we only compare which model is “smarter,” we may forget to ask whether we can leave when its terms become unacceptable.

Those omissions may reflect commercial messaging. They may also reflect media incentives, technical barriers, and our own preference for convenience.

Without evidence, we should not attribute every gap in public discussion to a coordinated effort by powerful interests.

But we do not need proof of manipulation to recognize that the missing questions deserve attention.

The Goal Is Not Panic. It Is Preserving Choice.
Huang’s departure does not, by itself, establish that the open-model ecosystem is coming to an end. Companies’ actual policies and decisions must be watched and assessed.

But this moment can still remind us of something important: public interests this significant should not depend on any one executive’s continued support.

We can appreciate good products and support legitimate commercial returns while also asking public institutions to avoid dependencies they cannot escape.

We can demand explanations for consequential changes. We can support genuinely usable open alternatives.

The goal is not to force everyone to use open models or to prohibit closed services.

It is to ensure that, whichever model of development succeeds, people retain the ability to compare, question, switch—and, when necessary, say no.

As AI becomes more capable, we should not ask only:

“How much can it do for us?”

We should also ask:

“When we disagree with the company behind it, how much will we still be able to do without its permission?”

[ZeonArka @ZeonArkaym]

A billing dispute that ended with my account suspended

I participated in a Google Cloud trial or promotional offer. I did not actively use the services, and months later charges were automatically deducted from my linked card.

I asked Google to stop the services and refund the disputed payments. The case was closed because I submitted it from a different Google account rather than the one linked to the charged card.

The account associated with the charges had also been suspended. After I regained access through an appeal, I found the instance from the promotional period still running. I immediately shut it down and released it.

Despite my earlier request, my bank notified me of another automatic charge of USD 98.27 on November 1.

After applying for unblocking and retrieving the account linked to the bank card,I replied to the payment notification and also contacted Google Cloud Support at [email protected] from the charged account, asking for a written explanation of the charges, a refund of the completed payments, and confirmation that no further charges would accrue.

I did not receive a substantive response. Instead, the Google account linked to my card was suspended.

The three amounts involved are USD 100.58, USD 115.33, and USD 98.27.

To be clear, I did not refuse to pay for the services Google provided, and I am not demanding a free Google Cloud service. I am asking Google to explain:

Why charges continued after I asked for the services to be stopped.

What the disputed payments were for.

Why the account used to communicate with Support was suspended, rather than my question being answered.

Google widely promotes a commitment to transparency and fairness. I am posting publicly because the billing dispute has now been closed on both the email and account channels, and I do not know how else to have this reviewed by a human.

If Google is willing to address this in writing, I am happy to share the details privately.