It was definitely more feelgood and productive in terms of inference sharing and directions. Or may be that has to do with humility and conviction figures like Kailash brings into the room.

Some of the key takeaways for me were the signals for good builders which initially were about the building skills now are about the distance from the problem as well along with technical foundations to mitigate risk and harm.

It's not like the session was particularly novel for it's ideas. But was definitely more tone setting, like most conferences anyways are.

The session now ends for Tea break.

Shared and disturbed infrastructure amongst the like minded organisation is a design problem that is actually not expensive to solve at all in terms of funding.

Software was always a black box, but now having non-technical people building something may not have the same trust, making the human behind the system even more important.

To another question, Kailash reiterated that just because you can, you should not build without understanding and considering risks and harms. Shared technical infra is the only solution and it has always been.

One question, a visually impaired prommager asked where will the accessibility approach from retrofitting it for universal needs go? Kailash reiterates the optimistic future of shared engine make it more possible for more builders making it more universal.

Fractional product personal is a possibility now helping ten orgs. Kailash is asked to and distills his obvious prediction with logical deduction that shared software generating engine with outer shell is already happening.

Undo the term capacity building, replace with product thinking skills, is Kailash's call, to be closer and intentional about the changing and required times.

Capacity building is a cliched washed out term. Real capacity building now goes beyond usual generic tools. Grounded small team with tech fundamentals and close to problem. None of this requires big funding, but good intent.

The way things are changing, Kailash says this hunch is that the robust engine in a generative software can change itself, could be the reliable solution.

Vibe coded prototype is not a technical solution, as Kailash says. Unless the nuances and vocabulary, the articulation is going to be skewed.

Kailash shares that their is no yet framework of the shelf to run with, but some common sense for sensitivity around data protection can be applied to be exceptionally careful and build guardrails.

Akhila shares a story about a tool built by a domain expert, but all the problems about the maintainance and data collection and possible harm about sensetive data.

For a non-techy, who can also churn code now, being close to the problem is crucial. Domain expertise with possible harms is the better mental model.

For technological impact in the social sector, the engineering skills and product and ux etc skills, those were originally periferal to engineering and also are in the centre.

No software is perfect and it will need the maintainance. Even with LLMs, beyond certain complexity, you will a need a human oversight that also takes the ownership of sanctity and correctness.

Shared infra and open source. Social dynamics that made the open source are also changing. Incentives that drove this open source is changing now to in the abundance of the code. Longevity and importance are the factors, even the jury is still out and it will take time to settle.

The risk and complexity is a spectrum. But risk is hardly being talked about. Correctness and robustness is important. Since, there is so much being built, correct things have value. Building same things in slightly different risky way could emergent pattern.

Building is not the tough part anymore, but what to build with judgement is ever more important since anyone can build.

Kailash gets asked about his predictions now in hindsight. He says the prediction are obvious because they are also rooted in deduction. Kailash leads the way to the changed fundamentals from first principals.

Year 1 was about importance of open source in social sector. Year 2 was how/what do we build. Now this Year 3 more about the impact of AI in social sector. OASIS is gets a mention, an tech+social sector alliance.

Kailash Nadh now has a conversation with Akhila. There is a theatre and music pushed and feels cringe in Kaikash's light remark.

The whole team is invited to the stage, whihc appears to be almost 25-30 people. The distilled version is that the team wants to create building blocks for the social sector to make their own processes and tools.

Two of the co-founders of the T4GC, almost in rehearsed theatrical and scripted poetic. Open. Build. Break. Listen. Experiment. Share. Enable. Team. Are words that anchor the chapters in this duet poetic prose.

Started. This is the third year. They aspire to turn in Tech for Social Sector from their own community gathering. The day is a not very packed. One of the post lunch session is cancelled due to speaker unavailability.

I see a list of organizations mentioned for this summit on the dias. Tata trust, APF, RNP, dasra, Rainmatter, Samagata are some notable ones mentioned.

The speakers and the attendees are filling the Audi now. Things definitely have taken traction suddenly.

The tote given has cheap but useful, simple and creative props. A jargon ticker whenever a speaker says one of them. Also a wide network log to find and connect with diverse people and zine about T4G Community and Idlistack.

There couple of stalls outside. Idlistack and Impact. They both are cathering to the NGOs looking for solutions to ease their deployment and server management solutions.

Well, I got here after the registration and coffee served outside for the attendees, now inside at the auditorium at 9:53. But outside appears more happening. No signs that the event will start in about five minutes.