Satirical cartoon "Museum of Meaningless Metrics" mocks "Tokens Spent" alongside legacy software development metrics
Gary Marcus shared the cartoon to mock token consumption metrics.
Positive users praise the cartoon mocking "Tokens Spent" as a meaningless metric, while negative users defend the metric as objective or criticize the agile analogy.
No Digg Deeper questions have been answered for this story yet.
Most Activity

@HazzimIO @carloscastrodev Story points is probably the most useful out of all of these honestly lol
You can also add 'Number of papers published'.
No comments.

@HazzimIO 🤣

@HazzimIO story points was a humiliation ritual

@HazzimIO Why not bugs fixed, features implemented?

@HazzimIO Story points are easily the worst metric considering how they are totally subjective and exploited.
Pull requests are a runner up because, of how the code varies but, usually there are significant standards for devs and pull requests

@HazzimIO You missed this:

@DanielStormApps @HazzimIO They were for me, until I worked with a high performance team that just shipped and it is been years since we use any story points, we use some sort of T-shirt sizing now just to gauge and prioritize but that is about it

@HazzimIO Is there one metric that actually works? I mean, all metrics seem to be approximations only. Even things like income says more about market conditions and lobbying than wheather the product is good or the employees work hard.

@HazzimIO Fab! I put this in my newsletter, Marcus on AI.

@HazzimIO related to Meaningless Metrics here’s SNL’s funny skit about Confusing Metrics
https://youtu.be/JYqfVE-fykk?si=yHhWtols_8MI9KTG

@PilsZehn @HazzimIO Revenue/employee or something

@carloscastrodev @HazzimIO I mean, that’s the same. It’s all just vibes. How many vibes do you think this would take? We average this many vibes a week. Tying days or whatever to points is where people get it wrong.

@KirillGoldinBiz @HazzimIO There are many reasons - Software engineering is a lot more spray and pray and experimental so failure rate is high - Power laws seem to exist. One success can pay for a lot of failures making measurements even harder - Complex dependencies making credit assignment hard

@HazzimIO One more to add: "hours in meetings"

@DanielStormApps @HazzimIO @carloscastrodev Absolutely not. 💀💀💀

@HazzimIO Lines of code? You misspelled 'code coverage'

@richardsontools @HazzimIO What was story points?

@KirillGoldinBiz @HazzimIO Yes and large companies increasingly remove autonomy from managers to make this judgement rendering the function itself somewhat moot over time. Attempts to introduce objectivity like Leet code etc are increasingly low signal.

@HazzimIO What do you mean Pull Requests? There's only one pull request after squashing?