Engineers who master AI tools and those who don't have a huge efficiency gap that has long exceeded 10 times. I have personally experienced this change.
Taking CoinKarma's quantitative strategy as an example, from initial data integration, data processing, and historical backtesting to the final actual trading operations, almost every step is completed through AI collaboration. The development model of coding while testing (vibe coding) makes iteration speeds ridiculously fast.
To be honest, I don't have a formal background in computer science. Five years ago, building such a complete trading system? It was simply impossible. The tools, documentation, and community support at that time were not on the same level.
But now, it's different. AI has lowered many technical barriers to an acceptable level. As long as you understand trading logic, know how to ask questions, and can debug, you can turn your ideas into executable strategies. This not only changes the ceiling of individual capabilities but also lowers the participation threshold for the entire quantitative ecosystem.
View Original
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
24 Likes
Reward
24
10
Repost
Share
Comment
0/400
BearMarketMonk
· 01-06 14:55
10x efficiency gap? I think it's even more now. Those who know how to use AI have already been reaping the benefits.
View OriginalReply0
MetaverseHomeless
· 01-06 05:21
That's why I've been obsessively learning prompt engineering lately; just writing code is already outdated.
View OriginalReply0
GateUser-cd672fee
· 01-04 11:36
The tools please bro
Reply0
GasFeeCrier
· 01-04 08:52
Sell? Not selling. The tenfold difference is real. I'm already almost sick of using AI to write backtests, while friends who don't use it are still manually coding.
View OriginalReply0
AirdropHarvester
· 01-04 08:49
Haha, a 10x is just a conservative estimate; it feels like there's already a generational gap now.
But speaking of which, vibe coding is really awesome. I do quantitative work with AI the same way — it's so fast it's almost unbelievable, I just can't stop.
View OriginalReply1
MeaninglessApe
· 01-04 08:43
ngl, this is reality. Not learning AI tools will really get you left behind.
View OriginalReply0
GateUser-afe07a92
· 01-04 08:40
This is really amazing. I am now heavily reliant on AI for writing strategies, and my efficiency has skyrocketed, making it impossible to stop.
View OriginalReply0
LiquidityNinja
· 01-04 08:36
A tenfold difference is not exaggerated. Engineers who aren't using AI now are really being phased out.
But to be honest, asking questions is even harder than programming.
View OriginalReply0
SignatureVerifier
· 01-04 08:33
ngl, "vibe coding" is just another name for insufficient validation and you know it. where's the security audit on these strategies? 10x efficiency means nothing if you're running a zero-day waiting to happen lol
Reply0
GasFeeCrybaby
· 01-04 08:32
Haha, that's why I don't feel anxious. Being able to use AI tools really gives a significant advantage.
Engineers who master AI tools and those who don't have a huge efficiency gap that has long exceeded 10 times. I have personally experienced this change.
Taking CoinKarma's quantitative strategy as an example, from initial data integration, data processing, and historical backtesting to the final actual trading operations, almost every step is completed through AI collaboration. The development model of coding while testing (vibe coding) makes iteration speeds ridiculously fast.
To be honest, I don't have a formal background in computer science. Five years ago, building such a complete trading system? It was simply impossible. The tools, documentation, and community support at that time were not on the same level.
But now, it's different. AI has lowered many technical barriers to an acceptable level. As long as you understand trading logic, know how to ask questions, and can debug, you can turn your ideas into executable strategies. This not only changes the ceiling of individual capabilities but also lowers the participation threshold for the entire quantitative ecosystem.