build2learn-pivot-ideas.md

Build2Learn.in — Pivoting in the Age of AI

The Catalyst

"Code is cheap. Show me the talk." — nadh.in/blog/code-is-cheap

Software development as we knew it is changing. LLMs have made building easier — anyone can spin up a clone of anything. So what matters now? Ideas, marketing, go-to-market, and real-world problem solving. This sparked a discussion among Build2Learn's core members on where to take the community next.

---

Ideas from Core Members

🏢 Aravind — Real-World Work Experience for Students

• College students have no idea how an employee works inside an organisation
• Instead of random hackathon teams, group students with experienced folks
• Give them the actual experience of building within a team/org structure

🧠 Bhavani — Unconference + Ideation + Problem Solving

• Pick hot topics (e.g., AI Agents) and discuss industry-wide patterns, problems, and practices
• Come up with solutions and publish them
• Goal: Produce something foundational that the rest of the world can use

📣 Suresh — Marketing & Launches

• With AI, anyone can build — the differentiator is marketing and sales
• Devs need to improve their marketing skills
• Build2Learn should do more launches — both from organisers and participants
• Ideation sessions can be done online (Google Meet etc.)

🎤 Aravind — Product-Focused Speakers

• Invite people who have built successful SaaS products to share insights
• Cover the full journey: idea → execution → go-to-market strategy
• While other communities focus on tech-based speakers, Build2Learn should focus on product-based speakers

👍 Rhikshitha

• +1 on the marketing/launches direction

📅 Shiny

• Suggested a GMeet to discuss further

---

Emerging Themes

1. Code is no longer the moat — ideas, execution speed, and distribution are
2. Bridge the gap between building and shipping (marketing, sales, launches)
3. Real-world mentorship — pair students with experienced professionals
4. Publish & ship — don't just build at events, launch and market what you build
5. Product thinking > tech talks — speakers who've shipped, not just coded

---

Collected from Build2Learn core team discussion, 1-2 March 2026

Build2Learn model (come together and build) is gaining heat, not because the meetup model (talks/workshops) went bad, or people grew bored of it.
But, because these big giants want us to use/pay for their tools
Everything from Anthropic, google, openai to vectordb.
The whole ecosystem is pushing people to build using their platform.

Don't learn agent coding - use n8n
Don't learn Data engineering - use hex
Don't learn frontend/backend - use lovable

Build2Learn, at its crux, is to make people feel empowered, that they can crack any technology, create and learn something new in a matter of 4 hrs

Current State
•⁠ ⁠Make the coolest thing I want using AI in a matter of 4 hrs
•⁠ ⁠⁠Learning(not technical)

What’s next?
From here, we can go in two directions broadly,

1.⁠ ⁠Like I said before, go deep, technically deep…. learn to debate process, methods, techniques, use case
2.⁠ ⁠⁠Like @⁨DSK Orangescape⁩ mentioned, make 'em all entrepreneurs. Help them ship, launch, collect money, and microsaas it.

But for the bigger question
1.⁠ ⁠What can we do that will contribute to nurturing the tech ecosystem/talent (people remembering build2learn between 2020-2024 was a sign we did well)
2.⁠ ⁠⁠What will empower the engineers for what's coming ahead?

Once we think through enough on our own,
Let’s set up a call and do a vision/mission exercise

I would like us to do a bit of all 2 directions Bhavani mentioned. Students should get the technical depth. Experienced people should get guidance to microsaas their product. So basically, not any one of these, but both as long as we run it for students and professionals.

Can we do themed events? Like, AI agents one day where everyone builds an agent.

Other formats

1. Let me introduce @thelearningdev's Colearning model. We take an open-source tutorial or a tutorial blog post(e.g., https://blogs.mayankpratapsingh.in/chapters/speech-to-text-from-scratch), follow it together, learn together, and help each other. There will be one facilitator, who is not a trainer(not workshop) but one step ahead of the crowd, ie, has tried the tutorial before. We can also bring an expert to share the knowledge part, keeping hands-on as co-learning
2. We can do a PapersWithCode session. 30 mins of paper reading, 2 hrs of trying to implement them together as groups

✅ Things Build2Learn CAN Do

Event Format

• Run full-day events instead of half-day (half-day isn't enough time to finish anything meaningful)
• Run events once per quarter as the baseline commitment — this is realistic given venue constraints
• If someone else wants to run additional events (bi-monthly or monthly), they can take ownership and make it happen
• Use an invite-only format — pre-select teams and ideas, which also increases attendance commitment (people feel selected, so they're less likely to cancel)
• Pre-build teams before the event — participants register as a team rather than forming random teams on the day. we can keep it hybrid. people who are registering as a team can choose "I have a team" + team name + other team members register with same team name. people who dont have a team will choose "i dont have a team. need help"
• Vet ideas beforehand — only accept ideas that can realistically be completed (or reach a meaningful prototype) in one day
• Require projects to be micro-SaaS or small-scoped — solving one specific, small problem (like the email sync tool or the course completion tracker)
• apart from quarterly event, unconference style monthly online events can happen. combination of how to ideate and formulate an idea, building open source project, launching, side project demos.

Value Proposition & Focus

• Focus on developers, students, and freshers — people who still need to understand code
• Enforce a "no vibe coding" rule — participants can use AI for searching, planning, and reference (like Google), but cannot use AI to generate code directly
• Brand this as "artisanal coding" / "handmade code" — a developer-focused event that could become a marketable movement
• Differentiate from other communities by being the place where people actually learn to build, not just generate demos

Infrastructure & Tech

• Build a WhatsApp communication tool (using open-source APIs) to personally reach out to participants before events

Venues

• Use the Saama venue connection — getting a venue there is not a problem
• Explore college venues — they come with computer labs, ACs, and internet; Devasundari ma'am (VP at a women's college) has connections and has sent students before
• Build long-term venue relationships (like how Chennai Python group has IMSc/Ramanujan College as a permanent venue, or like the earlier relationship with Entrans)

Operations & Communication

• Personally text every registered participant one week before the event to confirm attendance and free up spots for others if needed
• Automate manual tasks — banner/image generation tool is built; use the task management platform (discovered via Sangeet at the Stable event) for onboarding new volunteers
• Formalise decisions and share them in the group — if anyone wants to do things differently, they need to take ownership
• Let the market decide — don't force the community if there's no organic support; don't take pressure or tension over things not happening

Community Philosophy

• Be patient with volunteers — people have personal battles; don't push those who are going through tough times
• Whoever is doing the work gets to make the decisions — you don't get to have opinions on things you're not actively working on
• Outlast competing communities without burning out — most communities die within a year

---

❌ Things Build2Learn CANNOT Do

Event Format

• Cannot run monthly events reliably — venue hunting in Chennai is too difficult and competitive to sustain month-on-month
• Cannot expect half-day events to produce finished products — 3 hours is not enough; projects end up incomplete
• Cannot expect stranger teams to continue projects after the event — even teams with people who know each other don't follow through; 90-95% go silent in group chats
• Cannot run open/free-for-all events and expect high attendance commitment — people register and drop out last minute with no accountability

Venues

• Cannot easily offer value to venue partners — Build2Learn's freestyle format doesn't align with corporate goals (unlike Google/AWS-specific events that benefit the host's engineering team)
• Cannot negotiate strongly with venues right now — not in a position of leverage
• Cannot fully prevent college venues from misusing the partnership — colleges will use it for admissions marketing, claim the event as their own, and force students to attend (which kills morale and makes it feel like a mandatory lecture)
• Cannot guarantee internet/signal at every venue — even good IT parks have dead zones - will become 4G for example in Saama

Technical Enforcement

• Cannot selectively block AI code generation while allowing AI search at the network level — the same APIs are used for both, and HTTPS encryption means you can't inspect the traffic content
• Cannot do deep packet inspection with a Raspberry Pi — that would require a much more robust (and expensive) machine, plus there are SSL pinning issues on mobile
• Can only block at the DNS level — which means blocking entire domains (like Claude, Gemini, etc.), not specific features within them

Communication & WhatsApp

• Cannot use WhatsApp Business API cheaply — it costs ~₹1,000 per event, requires pre-approved templates, and you must initiate the conversation first (new rules as of Oct 2025)
• Cannot freely customise messages through the API — only template variable values can be changed

People

• Cannot force people to care about a project they don't feel ownership of — only the person with the problem will drive the solution

following,,