@kmr_dilip Hi Dilip - we resonate with this thought. That is why we are building CellVerse AI - https://t.co/EY4SXmq7IR. A complete ecosystem for Healthcare AI. Like to connect with you. Possible ?
"AI Agent" is a BUZZ word right now.
The tech is not ready yet.
AI agents are not possible without cheap reasoning models.
Truth is, 98% AI agents are built with just IF or ELSE logic with tool calling and RAG capability.
So they are not AI agents, but coded workers.
Here's all you need to know about AI Agents:
I worked with 40+ companies in 2024 and the most in demand product was - AI Agents.
But then I explained them What is an AI agent exactly.
An AI agent is a system that can think, reason, plan, execute and evaluate tasks on its own.
That system won't require human input for each step. It can learn from collected data and its own mistakes over time.
But is this Agent Tech available right now?
Unfortunately no, it's not production ready.
Because the main piece of Agent tech is missing - Reasoning LLM models.
Models like GPT-4o, Sonnet 3.5, or open source models like Llama 3 are not Reasoning models, but implementation models.
If you give them a detailed plan, they can execute perfectly 9/10 times but if you ask them to reason, think, plan and execute they mess up 8/10 times.
OpenAI o1 model showed glimpses of smart reasoning.
But it's not available to all via API (only tier 5)
and also its not cheap ($15 per million input tokens and $60 per million output tokens)
and lastly, it is slow (It takes 5 minutes on average to reason, think, and plan before executing and task)
So as of now, we can build an AI agent that can work on its own but the accuracy of those agents are bad. They need human oversight.
But in future we'll see 2 types of models:
1. Cheap and fast reasoning models
2. Ultra cheap and super fast implementation models
And AI agents will use a mixture of these agents for specific tasks.
How far is this AI Agent Tech? I think atleast 6 months from now.
AI agents also need better framework (large context window, better system than RAG and large number of tools in their armoury)
o3 model may solve reasoning issue and we also see some great open source reasoning models.
Agentic system is the biggest vertical opportunity in AI Tech.
Agents will learn, adapt and refine their logic on their own, and over time they'll become the best in their domain because of the collected structured data, and refined workflow. (but that will happen in future)
Bottom line: AI Agent OS (operating system) is not ready yet. It requires cheap, fast and better reasoning models, large context window, and better system than RAG.
It'll take some time to get there. But you can stll build:
- voice to voice agents
- research agents
- systematic agents
Build them for different b2b industries and sell them as a service or as a software.
Companies are pouring money into AI innovation to get ahead of their competitors. They need AI Growth Consultants and AI Tech Devs to help them build their internal AI stack.
Build these systems for them and learn about alot of different industries.
I predict that AI SaaS founders will build the best AI agents! If you're not playing with AI code yet, you're losing the game.
That's why I'm building @CodeGuidedev . It's turning out to be the home of AI coders.
We're just 17-days old startup with 956 AI coders building projects with AI.
CodeGuide 2.0 is 98% ready. (Mobile app, Community, Video Tutorials, Boilerplates, better Post-Download flows) ✨️
Join CodeGuide now and enjoy limited 30% discount on Yearly Membership. ✌️
Peace
CJ
@lenardmedia @nevmed@thesamparr@johnarrow Windows version is under works - will be launched soon. Join the windows Waitlist here https://t.co/KOZNxC6XMF.
MAC local download version is available for MAC users. https://t.co/elXtG5g0Mp
#Networking is a highly underestimated skill especially among engineers while I feel it's one of the top 5 skills that should be developed for success in today's business/ job world? Share your thoughts here...
Spending on AI in the Asia-Pacific region is expected to increase at a CAGR of 34% between 2020 and 2025. A major part of this investment is likely to come from organizations that have already adopted AI and are looking to implemen…https://t.co/I13tk0R1qp https://t.co/D8FWIfVtAd
CUSTOMER LIFETIME VALUE (CLV) Prediction Analysis Workshop for Professionals using Cellstrat Platform. It provides the ability to business professionals to understand their customer assets.
@machinelearning@CellStrat
https://t.co/zHwuH2bkbI
Bldg & Training Models is one thng & Deplyng 'em for the world to use is anothr. #AI Deployment has always been a challenging & complicated task, hence, introducing @CellStrat Hub API Platform. Chk this video in which #vision model is deployed as an API: https://t.co/ytCp77Rqro
Leveraging AI Forecasting Models in a Data-Light Environment — a Brief Guide
Demand forecasting is an integral component of several business functions within an organizational setup; from risk assessment to workforce planning all…https://t.co/NeiULUiJVe https://t.co/N5bpHAu7S1
#CellStratAILab#disrupt4.0 #WeCreateAISuperstars#MLOps#AI#ML
Presenting CellStrat DeepTech Jobs in AI, ML, MLOps and Data Science :-
Today CellStrat AI Lab is one of the most prestigious and largest AI communities in the World with 15000+ members, 5…https://t.co/gERvyPJlJy