Codementor Newcomer Post: "Decoding AI's Impact: Insights from a Recent Harvard Study"

"Start writing here... "

^ Maybe make this placeholder, not text? (looking at you, codementor frontend...).

Also, for some reason, my cover image got compressed to the point of pixelation.
Does anyone know why? :)

Hello World! My name is Max. 👋 I just joined this platform. Let me share something with you...

Are you ready for most recent Harvard research paper? (as of publication of this post)

(#personalpost, not an #AI post). I'll tell you why this is important a bit later.

Last Sunday I managed to find some discipline to read recent Harvard study about language model usage at work (quick link below):

🔗 https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4573321

This Harvard-led study revealed to me that:

• Chat GPT-4 boosts enterprise efficiency by 40%

• Significant performance improvements were observed among Boston Consulting Group (BCG, basically a really big firm in the US) consultants when using generative AI

• Lowest performers saw the largest gains with the aid of AI

• The research underscores the need for strategic deployment and collaboration between humans and AI.

Therefore, I'd like to share my conclusions and concerns, mainly looking at IT:

• I think this will make things complicated a tad bit in the IT market. Entry level jobs will be a strange place to be in for a while now. So is academia

• Junior developers should be careful with #GPT4 model usage (or any other AI models, like #LLaMA, #CodeLlama etc) when still in the learning phase. Yes, you will see more output, but it can tilt your personal progress. You can lose your authentic learning and skill development, be careful long term (!!!)

• Senior developers can benefit significantly by automating mundane tasks. This allows them to focus more on architecture, optimization, security, privacy and other critical aspects of system design, reducing routine and enhancing productivity. A breath of fresh air for seniors?

• Business can now comprehend the code base better (hopefully by hosting an AI model themselves or as a service from someone they trust). Simply conversing with the bot about what they need. They can "discuss" requirements directly with the bot, ensuring clarity and accuracy

• Sales teams can leverage AI-driven insights for personalized pitches and to predict customer needs, making every interaction more tailored and efficient

• A lot of people are right now mindlessly feeding their company data to 3rd party language model providers. Where is all the data stored and who has access? Is it well secured and not leaking?

• Waiting it out and seeing many AI tools come out that you can use daily is more productive than trying to create something of your own, unless you have a team or very specific requirements. There is no FOMO, imo.

• Enterprises upgrade what they already have. Startups live on, creating solutions now even better with higher clickbait.

Welcome to present moment! Future is unknown. You can read the study from Harvard yourself here (redirect to SSRN):
🔗 https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4573321

It's a big topic and things to talk about that can be endless. So I kept this clear, concise and easy to read.

Do you have any favorite open source AI projects? Or are you working on one?

What are your thoughts on authenticity during the rise of AI as a technology? I hope you guys enjoy my starter post on this platform. Feel free to check out my profile and message me about what you're working on.

Did you like the post?