$180k to Support Job Seekers in Bangladesh
With the University of Illinois Chicago (UIC), I built AI Leaders, a workforce development course. Our first cohort of 105 started in February: 35 interviews, 7 jobs at tech companies and universities. We've since grown to 1,820 learners, many in Bangladesh through BRAC University. Cost this year: $391,282, about $215 per learner.
The problem: UIC won't fund BRAC-related work. $180K covers an engineer, a community manager, and model and hosting costs for one more cohort, all in Bangladesh.
Why Bangladesh? AI will change how economic development works. My thesis: give people in developing countries the skills and experience to tackle the world's most pressing problems, and development spreads far more efficiently than through typical foreign aid.
Bangladesh is a great test case for that thesis because it's a developing country with talent, not the jobs. Graduate unemployment hit 13.5% in 2024, about 885,000 people (Dhaka Tribune). One in three graduates waits up to two years for work, and about 36% of job seekers find it through family and friends (Banglanews24). Our learners lack those connections, so we teach a structured search process: find roles, apply, and reach out to real people.
Why 11:11? 11:11 doesn't charge any administrative fees (my salary comes from UIC), so I can ensure 100% of funds go to the program. Alternatively, happy to bill expenses directly to a funder.
Give: Donate here to my 501(c)(3), 11:11 Philosophers Group, or email blake@philosophers.group to discuss other options.
If not cash, give your input: What's wrong with my theory of change? How should we tune further cohorts? The goal of this project is to raise wages in developing countries. How can we do that more efficiently?



One thing I find fascinating is how differently people in the same sub-field use it. One world leading cosmologist I know described LLMs a tutor in his field months ago. Another still doesn't use it at all. Many conversations with software developers over the course of the last year have yielded similar stories, although they all seem to heavily use it now. I was in awe of what my superforcaster friend could do with LLMs in economics when they were still consistently hallucinating in intermediate undergraduate economics for me. When he shared how good the instructions he'd put in his settings were I realized that was probably the cause of our different results. I've spent a lot of time working with LLMs in areas they still hallucinate tremendously in, so I hear you. I agree coders can get lost in their bubble. But my sense from my own experiences with LLMs and from some of these conversations is that using AI is a skill that gets a lot better with practice. It may take a large investment initially. I would naively expect LLMs to make a field worse in many ways at first but for that to improve substantially over time as humans get more skilled at using it, independent of improvements in technology. I would expect this to be true both on the individual level and the organizational level. What's your sense of how this capability overhang dynamic is playing out in your field?