Hugging Face's logo

Join the conversation

Join the community of Machine Learners and AI enthusiasts.

Sign Up
Back to feed
fdaudens 
posted an update 21 days ago
Post
3631
I read the 456-page AI Index report so you don't have to (kidding). The wild part? While AI gets ridiculously more accessible, the power gap is actually widening:

1️⃣ The democratization of AI capabilities is accelerating rapidly:
- The gap between open and closed models is basically closed: difference in benchmarks like MMLU and HumanEval shrunk to just 1.7% in 2024
- The cost to run GPT-3.5-level performance dropped 280x in 2 years
- Model size is shrinking while maintaining performance - Phi-3-mini hitting 60%+ MMLU at fraction of parameters of early models like PaLM

2️⃣ But we're seeing concerning divides deepening:
- Geographic: US private investment ($109B) dwarfs everyone else - 12x China's $9.3B
- Research concentration: US and China dominate highly-cited papers (50 and 34 respectively in 2023), while next closest is only 7
- Gender: Major gaps in AI skill penetration rates - US shows 2.39 vs 1.71 male/female ratio

The tech is getting more accessible but the benefits aren't being distributed evenly. Worth thinking about as these tools become more central to the economy.

Give it a read - fascinating portrait of where AI is heading! https://hai-production.s3.amazonaws.com/files/hai_ai_index_report_2025.pdf
ArefSy

The cost to run GPT-3.5-level performance dropped 280x in 2 years

This is insane

abiaselli

Comparing the cost of US to China is not super accurate considering currency differences and the fact CHIYNA does significantly more public investment. Also a decent number of those American papers are by Asian-Americans or actual Chinese natives.

paradiseDev

I think this is a little disingenuous to claim "the benefits arent being distributed equally"

You first rightly claim that the difference between open and closed source AI is vanishing, which provides more people worldwide with access to AI, and high quality AI, whether people have money or not,

then you say that since the US has invested 12x times more than china, and since more males have been paid than females in the AI job/skills market, that means benefits arent distributed equally?

You never provided the AMOUNT of males and females that actually work in the field, and I am willing to bet there is roughly a 2.39 : 1.71 male/female ratio..

and again, the two countries that are the heavest involved in AI, the USA and China, would naturally have the most money invested in it.. I fail to see the issue here.

are you suggesting we give some of the money that private investors are giving to the US on their own choice and will, to China instead? to "even the distribution of benefits"?

and if youre suggesting that places hire more women in AI just because they are women, then youre going to see the progress of AI absolutely cripple, not just because they are women but because they werent hired based on their qualifications and competency, they were hired because of their gender,

just as well, you will only see the gender wage gap increase if you hire more women because they are women, because it will be obvious right away that many new hires are not skilled or competent and they will be given lower level tasks and positions, while the truly competent and skillful will be paid higher and higher, to fix the unbalanced mix of skills and competency.

I fail to see the issues that you are trying to point out, and i think as i said at the beginning, you are being a little disingenuous about it.

its just data. it doesnt need to be fixed.

and not everything has to be equal.

In this post