The Cult of Microsoft
We recorded 3 claims from this post. 1 was checked against evidence. 2 can’t be checked. 11 more are held. Read the post at its source.
Claims from this post (3)
In the order they appear in the post. Each quotation is the speaker’s words in this post. A label such as “Supported” or “Can’t be checked” is the claim’s current assessment; its page has the evidence and the reasons.
Assessment: Can’t be checkedSaid
“Yet to actively encourage the company-wide dilution of performance reviews — and by extension the lives of Microsoft employees — by introducing generative AI is reprehensible.”
What we tested: Actively encouraging the company-wide dilution of performance reviews, and by extension Microsoft employees' lives, by introducing generative AI is reprehensible.
Assessment: Can’t be checkedSaid
“Third, lines of code is an absolute dogshit metric.”
What we tested: Lines of code is an extremely bad metric for judging software work.
Assessment: UnresolvedSaid
“At some point, these ugly messes will collapse as it becomes clear that their entire infrastructure is written upon increasingly-automated levels of crap, rife with hallucinations and devoid of any human touch.”
What we tested: At some point, these ugly messes (large companies that have given up on managing their people in favour of automation) will collapse, as it becomes clear that their entire infrastructure is built on increasingly automated, hallucination-ridden work with no human touch.
Held from this post (11)
These records did not pass every check, so nothing about them is published. Held is not a finding.
- A required check could not run. Our reading, not assessed: Microsoft asks employees to draft performance reviews from outputs of hallucination-prone generative AI models that also underpin ChatGPT.
- A required check could not run. Our reading, not assessed: According to a source Zitron talked to, Copilot's personalized achievement summaries are also prone to hallucinations.
- The two reviews read the claim differently
- The criteria are ambiguous. Our reading, not assessed: Microsoft instructs managers to use Copilot to summarize direct reports' Connects and feedback as a basis for drafting reward and promotion justifications.
- The record does not meet the current rules. Our reading, not assessed: Zitron is fairly confident that Google's AI is just writing comments and creating unit/integration tests rather than building key parts of its software.
- The two reviews read the claim differently
- The criteria are ambiguous. Our reading, not assessed: The difference between an AI model creating a class declaration and an IDE creating it is minimal.
- The two reviews read the claim differently
- A required check could not run. Our reading, not assessed: Microsoft is actively training its employees to use Copilot to generate their Connect performance review responses.
- The two reviews read the claim differently
- A required check could not run. Our reading, not assessed: When the Connect accomplishments prompt was run repeatedly, Copilot hallucinated wildly different statistics for the same metrics.