Do you have any concerns around scaling? I like LikeKit stack, but if not mistaken their agent architecture is based on multiprocessing (one os process per 'session'/'conversation') which doesn't sound very scalable. Btw, great demo, this is a cool technical problem to solve. I've spend a couple of months in this space (using a similar stack) and know for a fact that's not easy.
Thanks, there are certainly a lot of fun and challenging problems to take on in the space. On scaling, the agent architecture isn't limited to one machine, so you can also autoscale your machines. It's essentially python's Celery if you've tried that. It gets more tricky when you require GPUs though!
I didn't make this clear enough in the post, but we're still working on voice cloning and inflection transfer. Voice cloning is easier, but to support inflection transfer we have to modality-align an LLM.
I don't think you can get a mortgage with a 6% downpayment anymore. Canadian real estate is a scam. Banks would lend to anyone with a 5% downpayment. The mortgages are insured by the government (CMHC), so essentially all the taxpayers are on the hook if the real estate market implodes. Banks will be fine though as usual.
I've lived in both Bay Area and Toronto. Toronto has become absurdly expensive. You can get a 3-4 bedroom modern townhouse in Dublin/Pleasanton for 850K USD. It's common for a senior software engineer to make 200-230K. You can't buy an equivalent housing in Toronto for 850K CAD. Crappy townhouses in Richmond Hill go for 900K - 1M+ and similar housing will be 1.2 - 1.4M. it's also extremely difficult to make 200K+ CAD in Toronto (senior engineers make ~140K). A relative of mine does mortgages so has accurate statistics on how much people make.
