Will companies really prefer AI when it costs so much to reach human accuracy?

Seeking Advice
1

Just saw some buzz about OpenAI’s o3 launch, and people are speculating that AGI might already be here. But here’s the kicker: improving accuracy from 75% to 85% took 10x more in costs.

When regular employees are expected to hit 99% or even 100% accuracy, it feels like AI might not be as cost-effective as people think. Will companies end up spending more money correcting AI errors or just accept less efficiency?

Are we about to see a rise in bureaucracy with humans double-checking everything the AI does? Or could some industries realize AI is more expensive to operate than human workers?

2

Welcome to this forum’s AI discussions

Posting guidelines for better conversations

A few things to keep in mind when posting:

  • Keep posts over 100 characters, more detail helps spark discussions.
  • Your question might already have answers—use the search bar first to see if someone’s posted about it before.
    • For example, posts like ‘AI is going to take all the jobs’ come up often.
  • Feel free to share both positive and negative takes on AI, just keep it respectful.
  • Back up your arguments with links when possible.
  • No wild conspiracy theories, like AI summoning the apocalypse. It’s not happening.
Thanks, and let the mods know if you have questions or feedback!

I’m an auto-generated bot message. Reach out to the moderators if anything seems off.

3

AI is already valuable even without the o3 accuracy. It may not match human-level precision in every task, but it works way faster. Think about using AI for repetitive tasks so humans can focus on the critical stuff.

4

People won’t care about current costs. They’ll drop eventually.

5

AI doesn’t need breaks, never calls in sick, works weekends without complaints, and doesn’t ask for a raise or benefits. It doesn’t have off days, bad attitudes, or personal drama. No hiring, firing, or HR headaches.

Sure, it might not be as accurate as humans now, but its consistency and flexibility make up for it. Plus, AI could affect management roles as much as regular office jobs.

6

The short answer: companies will start with AI in roles where extreme accuracy isn’t critical, but the work is costly. For example, front-end developers.

Low-paying jobs like janitorial work, which AI struggles with, will stick around. High-paying jobs needing precision, like brain surgeons, are also safe for now.

Warehouse work and education are some areas where AI doesn’t need perfection to be useful.

7

Where are you hiring that employees hit over 90% accuracy? I’d like to meet them! :thinking:

8

Who actually expects 100% from any worker? Have you ever worked a job?

9

Why do you keep throwing out random numbers like they’re facts? Just curious.