@levelsio Really good! Also it’s worth noting that ChatGPT is super slow intentionally. Microsoft serve OpenAI models 5x faster on Azure. One would be missing out to use OpenAI hosted models through their API.
@OfficialLoganK Also, I use GPT-4 through Azure because it’s 10x faster. And you going to give Microsoft an edge forever? It’s fine, it’s your prime investor, as I know where to invest our time
@OfficialLoganK How to choose between fine-tuning and function calling to enhance a GPT with industry-specific knowledge (in my case it’s speaking Quebec French and having local gardening skills). Fine tuning seems eager to learn from Q&A data, function calling seem to provIde a slower UX.
@bhumi1102 Fractional CTO, CTO on-demand, Board member ex-CTO, tech-savvy investor, all the same. Point is you want a highly technical friend in your close circle of advisors doesn’t matter what business you are in. I really enjoyed that role and can help share insights if needed
After sitting with our legal team to parse through the fine print of Apple's DMA announcement (that took a while), which is, at best vague and misleading, I wanted to share my thoughts.
While Apple has behaved badly for years, what they did yesterday represents a new low, even for them. See more info here: https://t.co/2lznqwYujY