Is it more expensive to use AI than real people? Or vice versa? [Real costs]
Yesterday, I saw this report: Meta employees consumed 73.7 trillion AI tokens in a single month. This costs roughly $221 million a month and around $2.65 billion a year. That’s the salary of about 9,000 engineers.
In addition, the consumption of tokens (energy and electricity related to it) also has its limits.
Many people are starting to boycott AI. Today I read in a private message from a company profile: "Was about to sign up and noticed you use AI images, so decided not to"
I don’t doubt that AI makes our lives easier and creates opportunities in many ways.
But don’t you notice the fact that at a certain point it becomes unsustainable and disadvantageous?...
Where do you think the tipping point is?
How to eliminate the disadvantages that AI brings with it?
For me personally, the biggest downside is the loss/breakdown of trust.
Replies
Every company is experimenting at their own pace; $2.65 billion is pocket change for Meta.
Sundar Pichai's quote
"AI is moving so fast that humanity may struggle to keep up with its implications."
The tipping point will be when the production output volume outpaces judgement.
Its takes lot of effort to understand AI presentations or research reports generated in 5 min.
Especially when metrics are hallucinated and can't be trusted.
Eliminated the disadvantages of AI acting on its own like that sneaky extra git commit.
Humans think and AI executes. If we can't defend the AI output, maybe do not ship it.
minimalist phone: reduce your screentime
@roopesh_donde And maybe we just live in a bubble, because we are heavy in tech, but imagine the rest of population. 85% of people do not have any clue what is happening :)
@busmark_w_nika 100%, besides tech folks, the awareness is only surface level. A friend's role of "Financial Accountant" was eliminated due to restructuring, but he was unaware that AI is automating Desk jobs.
After watching 100's of podcasts from godfathers of AI and CEO's, I have realized that either we are in a I bubble, they way its marketed or things will drastically change in 5 years. Nevertheless we need to be involved to be relevant.
I think the cost debate misses the real shift, it’s less about AI vs humans and more about where trust breaks in the workflow.
In outbound, we built Dialbotix as an AI Voice SDR to handle the first layer of outreach and qualification. What we noticed wasn’t cost savings, but that teams stopped caring who made the call and started caring whether the lead was actually worth their time.
The tipping point isn’t economic, it’s whether AI is removing noise before human judgment, or replacing the judgment itself.
minimalist phone: reduce your screentime
@moeez1 yes, but hitting the right prompt requires many "bad prompts" before :)
the cost curve will compress over time, open weights, infra efficiency, competition. that part solves itself. the trust problem is harder because it's behavioral. once people pattern-match "AI output = hollow," they start applying that filter even to good work. that's the real tipping point and we're already past it in some audiences. the fix isn't less AI, it's using it where it actually adds signal instead of just adding volume. the companies getting boycotted aren't being punished for using AI, they're being punished for being obvious about using it badly.
minimalist phone: reduce your screentime
@rnagulapalle I think we test limits of AI, companies and humans at the same time :D
The Meta numbers are wild but I think they're the wrong reference point for most people. 73.7 trillion tokens across a company of 70,000+ employees is a centralized infrastructure cost — it's not what small teams or indie builders are dealing with.
For me the tipping point is simpler: does the AI output require less effort to fix than it saved you in the first place? When that ratio flips, you're losing.
I build a data report tool and use AI for executive summary generation. It works because I've constrained what the AI touches — structured input, specific output format, human review on anything customer-facing. The cost is predictable and the output is auditable.
The trust thing you mentioned is real though. I had a potential user back out recently because they saw "AI" in the product description without reading what it actually does. That's not a technical problem, that's a communication one — and no model can fix it.
minimalist phone: reduce your screentime
@deepanshu_garg9 to be honest. I feel it's a win for humankind when they rehire humans back because of overspending money on AI :D