Amandeep Singh

IIMA/L/C | M Tech | A Tech Enthusiast, Optimistic, Avid learner, relentlessly focused on what's next.

[Amandeep Singh @amandeep7s]

11/11

One lecture.

One mental model.

One question I can't stop thinking about.

Where will value accumulate as AI matures?

Curious how others are thinking about this.

[Amandeep Singh @amandeep7s]

1/11

Everyone is asking the wrong AI question.

The question isn't:

"Which model is better?"

The more interesting question is:

"Where is the money?"

That idea completely changed how I think about AI after watching Stanford's new course on the Economics of the AI Supercycle.

[Amandeep Singh @amandeep7s]

10/11

I don't have answers yet.

But I suspect these questions will matter far more over the next decade than arguing about benchmark scores.

[Amandeep Singh @amandeep7s]

@dtptraffic No improvement. Still pathetic traffic at the same place during evening hours. Stuck up from the past 30+ minutes.

[Amandeep Singh @amandeep7s]

@dtptraffic every day huge jam outside ISBT Kashmiri Gate. Stuck for more than 30 mins. Buses, autos halt there and occupy half of the carriage way. Why don't DTP control it?

[Amandeep Singh @amandeep7s]

After so much of investments gone into AI and everyone is talking about the benefits, is the year 2024 silently moving towards the next AI Winter? Will it lead to the Tech Winter?

#AIWinter #TechWinter

[Amandeep Singh @amandeep7s]

#Intel is trying to win back the AI game with the new #Gaudi3 AI chip. Intel claims it is 50% faster to train the #largelanguagemodel #LLM than #Nvidia H100. It is build on #TSMC 5 nm processor.

[Amandeep Singh @amandeep7s]

Create Security champions in the teams who prioritize security in decisions, raise concerns, and bridge gaps between dev and security teams. Trained by experts, they reduce risks and spread awareness. #SecurityChampions #SecurityMindset

[Amandeep Singh @amandeep7s]

Organizations are quickly adopting LLMs for tasks like content creation and decision-making. While they seem like a quick fix, it's important to consider if LLMs are truly the best fit. Many problems can be addressed more affordably with traditional NLP methods. #AI #LLMs #NLP

[Amandeep Singh @amandeep7s]

Think twice before fine-tuning Large Language Models (LLMs) for your needs. It can be costly and complex. Sometimes, simpler methods like Retrieval Aaugmented Generation (RAGs) work just as well and cost less. Balance your options before diving in. #LLMs #RAG #FineTuning

[Amandeep Singh @amandeep7s]

"Consistency enlarges ability."

-@JamesClear

[Amandeep Singh @amandeep7s]

Upcoming plans of Chandrayaan3
#ISRO #Chandrayaan3