cm0002 to Technology@midwest.socialEnglish · 23 days agoStarbucks quietly retired its AI agent just months after deployment after it miscounted coffee shop inventories and slowed down baristasfinance.yahoo.comexternal-linkmessage-square8fedilinkarrow-up127cross-posted to: technology@lemmy.ml
arrow-up127external-linkStarbucks quietly retired its AI agent just months after deployment after it miscounted coffee shop inventories and slowed down baristasfinance.yahoo.comcm0002 to Technology@midwest.socialEnglish · 23 days agomessage-square8fedilinkcross-posted to: technology@lemmy.ml
minus-squareWindex007@lemmy.worldlinkfedilinkEnglisharrow-up1·21 days agoI think that’s a fantastic use with a few caveats. Feel free to take this and start a consulting company on this: Is an LLM what I want? is the goal to parse or generate natural language? does your use case have tolerance for inaccuracy? If both answers are “yes”, go. Go full steam. Otherwise? No. Outgoing call center? Go. Reading resumes? Go. High tolerance for mistakes. Cast the widest cheapest net. Math? Health? Explosions? No.
I think that’s a fantastic use with a few caveats. Feel free to take this and start a consulting company on this:
Is an LLM what I want?
is the goal to parse or generate natural language?
does your use case have tolerance for inaccuracy?
If both answers are “yes”, go. Go full steam. Otherwise? No.
Outgoing call center? Go. Reading resumes? Go. High tolerance for mistakes. Cast the widest cheapest net.
Math? Health? Explosions? No.