Jay Dave v21.2

Aspiring AI/LLM Engineer

[Jay Dave v21.2 @davejay2003]

Capricorn ♑ Theme song

https://t.co/efMrZjgVMk

[Jay Dave v21.2 @davejay2003]

Open-source models, like those from Hugging Face, prioritize transparency and collaboration. They’re freely available for modification and local deployment, offering more control and flexibility.

[Jay Dave v21.2 @davejay2003]

Frontier Models vs. Open Source Models

In AI, frontier models and open-source models represent two distinct approaches to building and using large language models.

[Jay Dave v21.2 @davejay2003]

Frontier models, developed by companies like OpenAI or Google DeepMind, are cutting-edge and built with vast resources. They’re highly versatile and excel at general-purpose tasks, but they’re often proprietary, accessed through APIs or subscriptions, with limited transparency

[Jay Dave v21.2 @davejay2003]

@RobinSharma Pushups 💪🏻

[Jay Dave v21.2 @davejay2003]

@wojakcodes True, must break out of the risk adverse mindset for freelancing.

[Jay Dave v21.2 @davejay2003]

How do you organize your ChatGPT sessions? Let me know—I’m always open to learning new hacks!

[Jay Dave v21.2 @davejay2003]

I’ve been using ChatGPT actively for the past 6 months, and one simple tip has made a huge difference in how effective it is for me:

Create different chats for different purposes.

[Jay Dave v21.2 @davejay2003]

For eg, I have separate chats for coding, research, brainstorming, and general thought streams. This helps keep the context intact, so I don’t have to re-explain or lose track of ideas.

It’s a small adjustment, but it’s made my workflow much smoother and more productive.

[Jay Dave v21.2 @davejay2003]

I came across this fascinating YouTube video where a developer used AI to encode an entire codebase into ~400 lines, along with a ~250-line key to decode it.
Definitely worth a watch.
https://t.co/5GkEIj4x6x

[Jay Dave v21.2 @davejay2003]

MuSR (Multi-Step Reasoning): Tests how effectively an AI can handle tasks requiring multiple steps of reasoning.

MMLU Pro (Massive Multitask Language Understanding – Professional): Measures domain-specific knowledge across professional fields such as medicine, and engineering.

[Jay Dave v21.2 @davejay2003]

You must have felt the difference in how different LLMs are good/bad at certain tasks.
Like I feel ChatGPT is great when it comes to explaining concepts clearly, and Claude is more natural-sounding and witty in conversations.

[Jay Dave v21.2 @davejay2003]

MathLv5 (Mathematics Level 5): Evaluates the ability of LLMs to solve advanced math problems, covering topics like calculus, algebra, and beyond.

IFEval (Implicit Fact Evaluation): Assesses a model’s reasoning skills when dealing with implicit information.

[Jay Dave v21.2 @davejay2003]

The future of AI is not just about what it creates, but how it interacts with and controls the world around us!

[Jay Dave v21.2 @davejay2003]

You’ve probably heard by now that Agentic AI is the next big leap in artificial intelligence. But how is it different from the Generative AI we’ve been using?

[Jay Dave v21.2 @davejay2003]

The image illustrates the difference:
Generative AI is depicted as a creative brain, constantly generating new ideas and content.

Agentic AI is shown as a powerful, autonomous entity controlling various devices like smart homes, drones, and computers, taking action.

[Jay Dave v21.2 @davejay2003]

@kamath_sutra Build business solutions with LLMs