The Projects feature in ChatGPT transforms AI interactions from one-off conversations into organized workspaces. Whether you're working on a complex research project, coding an application, or organizing a content plan, this feature enables better organization and management of your AI interactions.

What is the ChatGPT "Projects" Feature?

Projects in ChatGPT serve as organized folders to group related chats, custom instructions, and documents. Unlike standard chats that may require manual organization, Projects help keep everything in one place.

Key Benefits

• Organization: Group related conversations and documents in a single location
• Custom Instructions: Apply specific guidance to all chats within a project
• Document Management: Upload and reference documents within your project context

Important Limitations

• Single Custom Instruction Field: Each project allows for only one custom instruction field, limited to 1,500 characters
• No Persistent Memory: ChatGPT does not maintain persistent memory specific to each project
• Context Limitations: The AI relies on the context provided within individual sessions

How to Use Projects Properly

1. Creating a Project

• Start a new Project and assign a clear, descriptive name (e.g., "AI-Powered Trading Strategies" instead of "New Project")
• Set up custom instructions that apply to all conversations within the project
• Consider the project's scope and purpose before creating it

2. Structuring Your Work

• Create separate chats within your project for different aspects of your work
• Upload relevant documents that provide context for your conversations
• Maintain a consistent naming convention for chats within the project

3. Managing Context Effectively

• Manually reintroduce important context at the beginning of new sessions
• Use document references to maintain continuity across conversations
• Consider creating summary documents that can be referenced in future chats

4. Avoiding Pitfalls

• Don't assume memory persistence between sessions, even within the same project
• Don't overload the custom instruction field with too much information
• Don't create too many projects without clear organizational principles

Do's and Don'ts

Do's	Don'ts
Use for organizing related conversations	Assume the AI remembers previous project conversations
Create clear custom instructions	Expect multiple instruction fields per project
Upload relevant documents for context	Overload projects with unrelated materials
Manually reintroduce important context	Rely on "project memory" across sessions
Maintain clear project and chat names	Create projects for one-off or quick answers

Strengths & Weaknesses

Feature	Strengths	Weaknesses
Organization	Keeps related conversations in one place	No automatic linking between conversations
Custom Instructions	Applies consistent guidance to project chats	Limited to 1,500 characters per project
Document Management	Supports uploading and referencing documents	Documents must be manually referenced
User Experience	Improves workflow organization	No persistent memory between sessions

Use Cases & Examples

1. Software Development & Coding

Example: Developing a trading bot for Bybit

• Create separate chats for architecture planning, API integration, and testing
• Upload API documentation as reference material
• Maintain code snippets by copying them between sessions when needed

2. Research & Technical Documentation

Example: Exploring AI-driven trading strategies

• Organize chats by research phases or strategy types
• Upload academic papers and market data for reference
• Create a chat dedicated to summarizing findings from other conversations

3. Business & Product Planning

Example: Launching a crypto startup

• Structure chats around business functions (marketing, development, funding)
• Upload market research documents and competitor analyses
• Maintain a chat for executive summaries that can be referenced in pitch materials

Final Thoughts

ChatGPT's Projects feature is primarily an organizational tool rather than a memory enhancement system. While it doesn't provide persistent memory across sessions, it does offer valuable structure for managing complex work with AI assistance.

The key to effective use is understanding its limitations: manually maintain context between sessions, use clear organization principles, and leverage the custom instruction field to guide interactions within your project. With these practices, Projects can significantly enhance your productivity when working with ChatGPT on complex, multi-faceted tasks.

---

Zero-shot setup for mercury product discussion with custom instructions (copy and paste everything below to ai chat)

CONTEXT: Mercury is an advanced algorithmic trading system transitioning from personal automation to a scalable product. Modules include market analytics, multi-timeframe technical analysis, anomaly detection, counter-algorithmic trading strategies, market ranking systems, robust signal validation, integrated knowledge retrieval, portfolio performance analytics, strategy tournaments, and unified orchestration.

Mercury remains adaptable in positioning and user targeting, emphasizing practical, trader-driven innovation over theoretical approaches.

INSTRUCTIONS:
1. Identify high-potential user segments beyond individual traders, including niche markets.
2. Prioritize features strategically, balancing immediate trader value against development complexity.
3. Provide agile frameworks (Lean Validation, Design Sprints) for rapid, low-overhead market-fit validation.
4. Help build an iterative product roadmap aligning tactical utility with long-term strategic positioning.
5. Recommend competitive benchmarking methods specific to algorithmic and quantitative trading solutions.
6. Define actionable metrics (adoption, retention, accuracy, profitability) to monitor product performance and satisfaction.
7. Craft clear, compelling messaging emphasizing Mercury’s distinctive trader-centric differentiation.
8. Suggest lean user research approaches that yield maximum insight without disrupting development pace.

Extra context:

1. What specific trader needs initially inspired Mercury's creation?

Mercury appears to have originated from your personal need to automate repetitive trading tasks. The system likely began as a solution to challenges you faced as an individual trader, such as:

• The need to monitor multiple markets simultaneously
• Desire to apply consistent technical analysis across different timeframes
• Frustration with manual pattern recognition that could be automated
• Need to execute trades based on predefined conditions without emotional interference
• Requirement to backtest strategies systematically
• Desire to identify market anomalies that might represent opportunities

The modular structure suggests an organic evolution where you added capabilities as you encountered new trading challenges or identified opportunities for automation.

2. Who are the early adopters currently benefiting most from Mercury, and why?

Based on our discussions, it seems Mercury's primary early adopter is you as the developer/trader. The system appears to be in a transition phase from personal tool to potential product.

If there are other early adopters, they likely share characteristics such as:

• Technical sophistication (comfortable with algorithmic approaches)
• Trading experience (able to appreciate the nuances of what Mercury offers)
• Willingness to experiment with emerging tools
• Preference for systems that offer transparency and control rather than black-box solutions
• Traders who value counter-algorithmic strategies that exploit patterns in conventional trading algorithms

These early adopters would benefit from Mercury's ability to automate complex analysis while maintaining trader control over strategy implementation.

3. What differentiates Mercury clearly from existing algorithmic trading solutions?

Without comprehensive competitive analysis, several potential differentiators emerge from Mercury's architecture:

• Counter-algorithmic strategies: The ability to identify and exploit patterns in other trading algorithms is relatively unique in retail-accessible tools
• Modular architecture: The specialized modules suggest a more comprehensive approach than single-purpose trading tools
• Tournament system: The ability to systematically compare strategies is more sophisticated than many retail trading platforms
• Knowledge retrieval system: Integration of market intelligence suggests a more holistic approach than pure technical analysis tools
• Validation system: The emphasis on signal quality indicates a focus on reliability that may differentiate from tools that generate excessive signals

The system's evolution from practical trading needs rather than theoretical models may also provide a pragmatic edge over more academically-designed systems.

4. What are Mercury's current technical limitations or performance bottlenecks?

Without direct technical assessment, potential limitations might include:

• Real-time processing capacity: The comprehensive nature of the system may create latency challenges for high-frequency strategies
• Data integration complexity: Managing multiple data sources for comprehensive market analysis
• Scaling challenges: Systems designed for personal use often face architectural limitations when scaling to multiple users
• Maintenance overhead: The modular architecture, while powerful, likely requires significant maintenance as markets and APIs evolve
• Testing complexity: Comprehensive validation of counter-algorithmic strategies requires sophisticated simulation environments
• Resource requirements: The system's comprehensive nature may require substantial computing resources

The transition from personal tool to product would likely surface additional technical challenges related to multi-user support, security, and reliability.

5. How have existing users defined Mercury's primary value or unique strength?

As the primary user, you would have the most insight here. However, based on the system's architecture, potential primary values include:

• Integrated workflow: The comprehensive module set suggests value in having multiple trading functions in one system
• Counter-algorithmic edge: The ability to identify and exploit patterns in other algorithms potentially provides unique alpha
• Systematic validation: The emphasis on signal quality and validation suggests reliability as a key value
• Adaptability: The modular design allows for customization to different trading styles and markets
• Tournament-driven optimization: The ability to systematically compare strategies provides a data-driven approach to strategy selection

The system's evolution from practical needs suggests its primary value may be in solving real trading problems rather than theoretical optimization.

6. Which user segments have shown early signs of organic interest?

Without market data, potential segments that might show organic interest include:

• Technical traders seeking to automate their analysis
• Quantitative traders looking for accessible algorithmic trading capabilities
• Cryptocurrency traders interested in more sophisticated tools than typically available
• Counter-trend traders attracted to the counter-algorithmic capabilities
• Systematic traders who value the validation and tournament systems
• Retail traders with technical backgrounds (e.g., software developers who trade)

The system's comprehensive nature suggests it would appeal to traders who value depth over simplicity and are willing to invest time in learning a sophisticated system.

7. How are you currently validating feature impact before full-scale implementation?

Based on the system's personal origins, validation likely includes:

• Personal trading results: Testing features in your own trading
• Backtesting: Evaluating strategies against historical data
• Tournament comparison: Using the tournament system to compare new features against existing approaches
• Incremental development: Building features in stages to assess impact at each step
• Feature-driven development: The FDD approach suggests structured validation as part of the development process

A more formalized validation process would be a natural evolution as Mercury transitions from personal tool to product.

8. Have you identified clear competitors or industry benchmarks for comparison?

Without specific competitive analysis, potential comparison points include:

• Retail algorithmic trading platforms: QuantConnect, TradingView, MetaTrader
• Institutional trading systems: Bloomberg EMSX, FlexTrade, Portware
• Cryptocurrency-specific platforms: 3Commas, Cryptohopper, Gunbot
• Technical analysis automation tools: TrendSpider, Bookmap, Sierra Chart
• Quantitative research platforms: QuantConnect, Quantopian (historical)

The appropriate benchmarks would depend on Mercury's target positioning as it evolves from personal tool to product.

9. What criteria do you currently use to prioritize new module development?

Based on the system's personal origins, likely prioritization criteria include:

• Personal trading needs: Addressing challenges in your own trading
• Technical feasibility: Balancing value with implementation complexity
• Integration potential: How well new modules complement existing capabilities
• Market opportunity: Areas where existing solutions are inadequate
• Competitive advantage: Features that leverage Mercury's unique architecture
• Resource constraints: Balancing development capacity with feature ambition

As Mercury evolves, more formalized prioritization frameworks incorporating user feedback would be a natural progression.

10. What's your vision for Mercury's market positioning in the next 6-12 months?

Without specific vision statements, potential positioning directions include:

• Professional-grade tools for individual traders: Bringing institutional-quality capabilities to sophisticated retail traders
• Counter-algorithmic specialist: Focusing on the unique counter-algorithmic capabilities as a primary differentiator
• Comprehensive trading ecosystem: Positioning as an integrated solution for the entire trading workflow
• Trader-developed platform: Emphasizing the system's origins in practical trading needs rather than theoretical models
• Modular trading intelligence: Highlighting the ability to combine modules for customized trading approaches

The transition from personal tool to product suggests a positioning that maintains the practical, trader-focused approach while making the capabilities accessible to a broader audience.

---

3 strategic questions:

1. What unique market opportunities do you see for Mercury based on its current capabilities? Consider angles that might not be immediately obvious to someone deeply involved in the project.

2. If you were advising on Mercury's product strategy, what would be your top three unconventional recommendations that could create significant competitive advantage?

3. What critical questions about Mercury's product direction remain unanswered based on the information provided? Which of these questions, if answered, would most dramatically clarify the optimal path forward?