This Wednesday, The Valley’s meeting space was filled with people from around the community. The subject? AI solutions to save you time and money. Andrew dropped a lot of great tips on how to properly prompt and get the most out of these tools. So lets dive into it.
If you’re more of a video person, here are the highlights from the evening:
Andrew started off by clarifying what AI largely means today: LLMs (Large Language Models). These are pattern matching engines that process human language. That said, they aren’t sentient or deterministic, meaning 1 + 1 might not always strictly equal 2, but rather “probably 2.” This probabilistic nature is core to how they work. Think of AI as an employee that can access and process vast amounts of information to generate responses.
Which One Should I Use?
While there are several key players like Google Gemini, OpenAI’s ChatGPT, and XAI’s Grok, Andrew recommended Claude (from Anthropic) for its all around great general model and a company ethos that appears more aligned with humanity. Also this technology is not as new as you might think. You’ve likely been using AI for years through simple Google searches.
Prompting 101
Like any good builder knows, the quality of your output depends on the quality of your input. Andrew laid out clear concepts for effective prompting:
Be Specific, Not Vague: Don’t just ask “What’s for dinner?” Instead, “Plan a week of dinners for 4, $80 budget, no seafood, using items from this Hannaford flyer (link provided).”
Provide Context: Basically all mainstream AI tools have “memory” that learns from your past chats. This allows them to tailor responses, remembering things like your kids’ food preferences or recurring maintenance needs. Use this to your advantage.
Define Your Format: If you want bullet points, a spreadsheet, or an image, ask for it. Treat AI like an employee: give clear direction. Although Andrew suggests with the more premium models, you can almost treat it as a higher level employee, and giving it more creative freedom and agency will result in even better results. So experiment and see what works best for you.
Iterate, Don’t Settle: Andrew’s go-to hack (aka shortcut), if you get an answer, simply respond with “Are you sure?” or “Double-check your sources.” This often prompts the AI to refine its output, boosting accuracy.
Leverage Multimodality: Text is just the start. Photos and audio are information-dense shortcuts. Snap a picture of your fridge, record a weird car noise, or upload a confusing bill. AI can interpret these visual and auditory inputs remarkably well. Just make sure what you are uploading is safe, because with non-local models it’s probably going to be used as training data.
Ways to use AI
Andrew covered a hundred use cases, but several categories stood out for their immediate, tactical value:
PS. Not kidding about the 100 use cases. Here is a PDF of the slide deck
Smart Meal Planning: Combine a fridge photo with a grocery flyer link to generate custom, budget-friendly meal plans that reduce waste.
DIY Diagnostics: Before calling a contractor or mechanic, use AI. A photo of router lights, a video of a strange engine noise, or a picture of a plumbing problem can guide troubleshooting and help you communicate more effectively with professionals.
Budgeting & Finance: Upload last month’s bank transactions for an instant budget analysis. Decode complex medical bills or compare insurance policies, as Andrew did to find a better family plan.
Educational Support: Generate homework help, flashcards, or study guides tailored to specific subjects.
Travel & Planning: Find cost-effective travel options (Andrew found a flight for half the price by shifting airports) or plan family weekends on a budget.
Wellness Routines: Create personalized workout plans based on available equipment and fitness goals.
The Faux Interviewer (One of Andrew’s favorites): Upload your resume and a job description. Instruct AI to act as the hiring manager. Repeated practice builds confidence and sharpens responses, preparing you for the interview process.
The Trade-offs
Andrew didn’t hesitate from mentioning the real trade-offs of this technology. He highlighted the ethical void in AI’s current model: its lack of compensation for the creators whose data it’s trained on. This is where legislative action is truly needed, rather than on broader, less-understood controls.
Furthermore, the shift from traditional SEO to “Answer Engine Optimization” (AEO) demands a new approach for businesses. Being the single, most comprehensive, and well-structured answer to an AI query is now crucial. Local businesses need strong Google Maps listings and detailed, Q&A-style content more than ever.
Recap
AI can’t do everything, but it’s a powerful set of tools that, when understood and used correctly, can significantly increase our capabilities. It’s a system to be learned, like any trade. Andrew’s talk wasn’t just about showing off; it was about empowering us to integrate these mechanics into our daily grind for more efficient, authentic lives and businesses. You can experiment with it, and leverage its power to build better systems.
Thanks to everyone who joined us. This event was in partnership with Cambridge’s Mud U initiative. Thank you to the Mud U team for all they do for the community.
Until next time,
-Charles Dill