If the firms literally marketing AI expertise to your board are the ones publishing AI hallucinations as research, that’s not a competence problem. That’s a signal that nobody actually knows what they’re doing yet, and the confidence is completely decoupled from the capability.
They’re not incompetent. They’re just running the same playbook as everyone else: ship something that sounds right, iterate when caught, move on to the next client.
source: https://www.ft.com/content/7e149ac8-2ce2-4266-8940-192f9821b33c?accessToken=zwAAAZ-saNlskc9-FJrILOJCZtOJQBkvmCGzPA.MEQCIB4QcO-EieUBJcSm2aKVE-2CPlzIp6X2ybz0z7CtBCWVAiBW2x-m3L_HC80hcP_ZAaHnRwPU6o3Lsz5ykpMfPd5r8g&sharetype=gift&token=6a7e4856-f5f9-4071-996e-8fedf70334e7
Feels like the QA version of this is when we shipped that clip-through bug because we were "out of time. " Nobody was incompetent, the deadline just ate the confidence check.
Kinda wonder if consulting firms have the same problem, just with billable hours instead of ship dates.
