RA_Fisher

9 ms·

151.
by RA_Fisher 3y ago
I’m curious, how do y’all keep track of performance and reliability? I ask, because I think it’s going to be a big challenge, so I built a service to record feedback / acceptance data: https://modelgymai.com/ If you th
152.
by RA_Fisher 3y ago
Who has higher inflation?
153.
The Impenetrable Autistic Glass Ceiling: Rethinking Workplace Inclusivity(medium.com)
4 points by RA_Fisher 4y ago|2 comments
154.
Artificial Specialized Intelligence: The Changing Landscape of Intelligence(datawonk.notion.site)
2 points by RA_Fisher 4y ago|1 comments
155.
by RA_Fisher 4y ago
As AGI economizes general human intelligence, people will scramble to acquire relatively specialized intelligence since it is now much more scarce relative to general intelligence (and both are highly demanded).
156.
by RA_Fisher 4y ago
I don’t think the same is as true of LLMs unless they’re put into a system (that can acquire knowledge and skills).
157.
by RA_Fisher 4y ago
Things like AutoGPT and BabyAGI seem intelligent to me. Intelligence is defined as the ability to acquire and apply knowledge and skills. How could it be denied that things like AutoGPT can acquire knowledge and apply skills?
158.
by RA_Fisher 4y ago
That wishful thinking stood out to me, too. “High-quality Soviet …”
159.
by RA_Fisher 4y ago
They ostensibly noticed the bank was insolvent and shared that important information with their startups. The bank should never have been anywhere near insolvent.
160.
by RA_Fisher 4y ago
It requires exceptional executives and an exceptional executive data scientist today, no doubt. Though, based on my experience it is possible given mutual due respect. Executives are domain experts, and I’m a business economist-statistician
161.
by RA_Fisher 4y ago
Both are unfortunately very deadly and there’s a lot of societal value in preventing both (without inducing more costly trade offs). Covid appears to be killing at a rate of 138,000 per year in just the United States.
162.
by RA_Fisher 4y ago
I’ve been performing it informally for many years and more formally for the last year and a half as Zapier’s Principal Data Scientist. I really enjoy combining my advanced technology, business, economic and statistical skills with large sca
163.
by RA_Fisher 4y ago
I think it’d be wise for more companies to employ an executive data scientist.
164.
by RA_Fisher 4y ago
And the stock seems down 7% vs. the Nasdaq.
165.
by RA_Fisher 4y ago
I get to apply my causal inference and measurement skills to a lot of compelling situations in collaboration with some really intelligent and skilled people. I truly feel like I get to perform technically excellent work, and I feel very pri
166.
by RA_Fisher 4y ago
I think the recent industry layoffs reflect that in part. Though I think there’s a better way— that is executive data science, what I do at Zapier. The key is that I’ve built up a huge amount of econometric and economic / business skil
167.
by RA_Fisher 4y ago
Both suffered ‘08 and the pandemic and have responsible central banks.
168.
by RA_Fisher 4y ago
It doesn’t sound to me like they’ve put in the work to earn the ownership share. I think it’s unfair to yourself and remaining team to keep them around.
169.
by RA_Fisher 4y ago
We mostly type these days and block writing’s good enough, so it’s a waste of time for the masses in my view.
170.
by RA_Fisher 4y ago
It’s nice, but it’s hard to beat the clarity and expressiveness of dplyr and purrr.
171.
by RA_Fisher 4y ago
That’s awesome customer service! Thanks for sharing.
172.
by RA_Fisher 4y ago
So true, causal inference is much more valuable than prediction which is a relatively cheap commodity.
173.
by RA_Fisher 4y ago
The person advertising has all the data they need to determine performance, just add statistics know-how.
174.
by RA_Fisher 4y ago
Yes! I love Apple Arcade for this reason, too.
175.
by RA_Fisher 4y ago
Exactly, RCTs take the mystery out. Nice work!
176.
by RA_Fisher 5y ago
The groups aren’t segmented by freeway vs. non-freeway use, but it’s an important omitted lurking variable. After conditioning on the variable, the takeaway changes.
177.
by RA_Fisher 5y ago
Yep. I wonder how realistic a 51% option would be?
178.
by RA_Fisher 5y ago
True. One would want to continually decorrelate services or model the dependencies. Redundancy will help even with some dependency, but you raise an important point.
179.
by RA_Fisher 5y ago
Bingo.
180.
by RA_Fisher 5y ago
True with two independent servers at 90% each, that’s 0.1^2 = 1% chance both fail— so redundancy can add a lot of reliability.

More ›