3 ms·
If you look into what happened recently with Deloitte and the Australian gov't (among many examples) the end product was so filled with errors and ficticious re
by LurkandComment 1y ago
If you look into what happened recently with Deloitte and the Australian gov't (among many examples) the end product was so filled with errors and ficticious references that it had to be reworked. It's like people using AI aren't selling you a solution, but the idea or confidence in a solution more so than we were before. It's a real extreme version of the value of a product is really a social agreement between the buyer and seller and not a monetary one. "I accept that this is what you say it is".
- ponector 1y agoBut any AI evangelist will say how massive performance boost they have with AI tools! Yes, people can make plausible crap much faster now. From my experience, since AI tools has been adopted (unofficially) by our developers, the amount of bugs increased dramatically. Feature delivery is much slower due to multiple rounds of testing after fixes.
- palmotea 1y ago> From my experience, since AI tools has been adopted (unofficially) by our developers, the amount of bugs increased dramatically. Feature delivery is much slower due to multiple rounds of testing after fixes. With AI, to realize the productivity benefits, you just have to adopt a process that create new stories for all those tasks. Then you'll start to see massive improvements when you look at your close rates. Just think of those new bugs you can close now that you're using AI!
- bespokedevelopr 1y agoThe answer to bad AI results from businesses won’t be, let’s forget about AI and instead hire back all those 6 figure programmers. Instead it will be just as it is has been for years, let’s send this (in this case) AI generated data to Malaysia or Bangladesh and have them clean it up and then bill it as perfect AI solution. These people in poorer nations have been cleaning data for ML and AI for years. It’s just adding another loop back in.