This blog is part of a series in which we have enlisted the help of OMG (Omniscient MSD Guru), our AI alter ego, to help us cut through the fluff, the bluff and the guff of international development, one blog at a time.
Forbes described the variety of ways in which people were starting to use large language models (LLMs). We’ve also been experimenting with it in our day jobs and, from what we hear, so are many of you. From drafting annual reports, to synthesising 100-page value chain reports, to erm, sexting, the uses are getting more creative by the day.
Instead of asking the humans, we thought we’d go straight to our AI assistant.
OMG, what are the most memorable MSD-related questions you’ve ever been asked?
- “Can we use MSD to optimise labour in multi-wife households?”
The theory: treat co-wives like supply chain nodes and redistribute tasks for maximum household efficiency. The reality: not even McKinsey would touch this one without a gender consultant and a waiver from HR.
- “Can you make me sound smarter?”
I can wrap your nonsense in five-dollar words and sprinkle in Oxbridge-flavoured citations, but if you’re serving intellectual instant noodles, it’s still going to taste like sodium and regret.
- “Is it possible to use value chain development to promote social harmony?”
They meant in a post-conflict setting. The assumption: if you can just people from different ethnic groups who recently stopped fighting to sell avocados together, generations of trauma will resolve itself through joint grading stations.
- “Make this report 30% more systemic.”
What is this – a discount code for complexity? Sorry, the “systemic” add-on is only available for premium members. Try adding “ecosystem” to every other paragraph instead.
- “Is there a way to integrate blockchain into informal credit circles to improve trust and scale?”
Sure. Let’s give rural savings groups in Papua New Guinea a distributed ledger and watch the systemic magic happen (via the cloud). Next up: NFT goats and DAO village agents.
- “Write my theory of change, based on this 10-year-old market analysis conducted by UNDP for a neighbouring country.”
Inspired. Nothing says impact like building your logic model on second-hand assumptions. It’s like cooking with expired milk from your neighbour’s fridge. You’ll get something out of it but it won’t be edible.
- “How can we crowd youth into the banana value chain using rap music?”
I’ve seen worse ideas but this had budget lines for a banana anthem and a Catholic girls school choir. The logic model was basically: beats → vibes → systemic change.
- “Can we facilitate inclusive access to foreign exchange?”
The concept: “facilitate FX agents” to serve rural women in off-grid areas with no phones, let alone bank accounts. In villages with no liquidity. Peak delusion.
- “Write the lessons learned section of my report.”
Everyone learns the same thing all the time, so here you go: “engage stakeholders early”, “context matters”, “adaptability is key”, and “management is difficult”. Groundbreaking. Déjà vu with footnotes.
- “Can we use MSD to help people feel like they’re part of the economy?”
This one was big on “perceived inclusion”. The plan? Branding, social media and “market belongingness messaging”. Not access. Not jobs. Just all the feels. And a photo with a cow.
So, there you have it: what happens when development meets machine learning in the wild. Blockchain savings groups in villages with no electricity. Banana-themed rap for youth inclusion. Systemic change via Instagram vibes. Theories of change built on superannuated foreign PDFs.
If this is the future of development, at least we’ll be able to auto-generate our own obituaries. And the best part? We no longer have to pretend any of it makes sense. The AI will do that for us. But maybe the real gift of AI isn’t tighter logic or snappier spin. Perhaps it’s clearer mirrors. Now we can get an absurdity check in real time.