Re:KPMG's AI report turns into a demo of AI hallucinations
by ahartmetz on Saturday June 13, 2026 @07:33AM
(#48516187)
Every report about AI problems has this comment. Just one more model, bro.
[ Parent
| on HN ]
Re:KPMG's AI report turns into a demo of AI hallucinations
by dijksterhuis on Saturday June 13, 2026 @07:38AM
(#48516236)
just one more model. just one more prompt. just one more guardrail. just one more. just. one. more.
Re:KPMG's AI report turns into a demo of AI hallucinations
by neonstatic on Saturday June 13, 2026 @09:31AM
(#48517174)
It's the 2026 version of "two more weeks"
Re:KPMG's AI report turns into a demo of AI hallucinations
by inigyou on Sunday June 14, 2026 @10:39AM
(#48527647)
Is that referring to COVID? The USA never started the clock on its first two weeks.
Re:KPMG's AI report turns into a demo of AI hallucinations
by neonstatic on Monday June 15, 2026 @01:45AM
(#48537058)
Yes, but not just COVID - it was the favorite fad around that time. When Trump lost the election there was a wave of grifters promising a reckoning to his fanbase with goal posts being moved roughly every two weeks.
Re:KPMG's AI report turns into a demo of AI hallucinations
by root-parent on Sunday June 14, 2026 @10:14AM
(#48527402)
AI is the new Fusion...
Re:KPMG's AI report turns into a demo of AI hallucinations
by gdhkgdhkvff on Sunday June 14, 2026 @11:35AM
(#48528339)
And every report about AI problems has this comment. Just one more failed use of AI and then AI will stop being useful and go away forever, bro.
Re:KPMG's AI report turns into a demo of AI hallucinations
by probably_wrong on Sunday June 14, 2026 @11:52AM
(#48528592)
I think you're misreading the parent comment.
As an example: Tesla has promised unsupervised self-driving for its cars for the last 10 years. Every year the promise is "it will be here soon", and yet this year we found out that the promises from 2019 are never going to materialize because the hardware isn't enough [1].
That's not to say that electric vehicles are going to go away, nor that supervised self-driving isn't useful. But it does mean that there are some things that the technology just can't do and that constantly arguing "the next version will fix it" without any evidence is not productive.
[1] https://www.motortrend.com/features/the-problem-with-tesla-u...