> My grandfather was an electrical engineer. He told me: do it well, or don’t do it at all. Simple rule. It guided how I built teams, how I shipped software,...
Not sure I buy that this perfectionist approach is the best in terms of ROI. Returns might be higher but investments disproportionately more so.
I hear your frustration about the quality of the work. But you also need to account for the way lower costs to reach this level of quality. If it's good enough, so what...
My understanding is that this is what the LoRAs are for; my belief is that they serve as "memory" to their live observations (a more NN-like cache, say), while the main LLM remains unchanged. These LoRAs are also weighted, so that LoRAs irrelevant to the current task will not be trained, while the relevant LoRAs will be reinforced.
But I never built it, so I am not sure if such an emergent state will appear or not.
“We’ve always encouraged people to have a break with KITKAT — but it seems thieves have taken the message too literally and made a break with more than 12 tonnes of our chocolate,” a KitKat spokesperson said in a statement. “Whilst we appreciate the criminals’ exceptional taste, the fact remains that cargo theft is an escalating issue for businesses of all sizes. With more sophisticated schemes being deployed on a regular basis, we have chosen to go public with our own experience in the hope that it raises awareness of an increasingly common criminal trend.”
Not sure I buy that this perfectionist approach is the best in terms of ROI. Returns might be higher but investments disproportionately more so. I hear your frustration about the quality of the work. But you also need to account for the way lower costs to reach this level of quality. If it's good enough, so what...
reply