Jishan Ahmed
@JishanCodes
Followers
38
Following
87
Media
40
Statuses
105
Building x. Helping you dominate productivity with AI
India
Joined September 2023
I replaced my virtual assistant with a Python script. It scrapes leads, qualifies them with @OpenAI GPT-4o, and drafts the email automatically. Cost: $0.12 vs $500/month. Here is the workflow logic: 👇
1
0
1
11) This project started from a personal experience, but it opens up deeper questions about AI and human-like interaction. If this concept interests you, feel free to reach out.
0
0
1
10) Looking for Engineers This is still an idea, but turning it into reality will need collaboration. Anyone with experience in NLP, deep learning, or AI memory systems who finds this interesting—let’s connect.
1
0
1
9) Ethical Considerations An AI that mimics real people raises ethical concerns. The challenge is ensuring that it remains responsible, transparent, and doesn’t cross into emotional manipulation.
1
0
0
8) Deployment & Cost Considerations Training and running AI models at scale is expensive. If this project moves forward, cloud hosting, API usage, and model inference costs will need to be optimized to keep it sustainable.
1
0
0
7) Technical Approach Some ideas for implementation include: Fine-tuning LLMs on categorized personal datasets Using RAG for contextual retrieval Experimenting with reinforcement learning to adjust responses over time
1
0
0
6) Predicting Future Responses A real personality doesn’t just react—it anticipates. A model like this would need to generate responses based not just on the present input but on patterns observed over time.
1
0
0
5) Handling Emotional Shifts Humans don’t instantly forgive or forget. One challenge is designing a system where emotions like anger or sadness don’t disappear immediately but instead fade over time based on interactions.
1
0
0
4) Memory & Context Retention An AI that can maintain long-term emotional context would require a structured memory system. Vector embeddings and memory decay functions could be used to simulate how people recall past interactions.
1
0
0
3) Understanding Personality A real person’s responses are shaped by past experiences, beliefs, and emotions. This project will experiment with fine-tuning models on categorized data that reflect emotional states and reasoning patterns.
1
0
0
2) The Core Challenge LLMs today don’t retain long-term emotional context. They reset after every interaction. This project is about exploring ways to integrate memory, emotional intelligence, and personality consistency.
1
0
0
1) The Idea Most AI chatbots forget past conversations and respond generically. The goal of Anu is to create a system that understands not just words but the reasoning and emotions behind them.
1
0
0
Building an AI that doesn’t just chat—it understands emotions. X9B7Q is an idea for an AI that remembers past conversations, reacts emotionally, and responds based on real personality traits. (A thread on the vision & challenges behind this project)
1
0
1