RA_Fisher

7 ms·

91.
by RA_Fisher 11mo ago
> On the fateful evening of November 10, 1975, McSorley reported he had never seen bigger seas in his life.[70] Paquette, captain of Wilfred Sykes, out in the same storm, said, "I'll tell anyone that it was a monster sea washin
92.
by RA_Fisher 11mo ago
“We’re holding our own” - Captain McSorley., 7:10pm, 5 minutes before contact was reported lost. Seemingly more a story of patriarchal arrogance than force of nature.
93.
by RA_Fisher 1y ago
For statistical AI models, we can use out of sample prediction error as an objective measure to compare models. What makes evaluating LLMs difficult is that comparisons are inextricable from utility (whereas statistical AI models do have a
94.
by RA_Fisher 1y ago
Or, maybe you don’t understand why it’s rational?
95.
by RA_Fisher 1y ago
Yes, for sure! And being able to orchestrate AI to use that knowledge provides leverage for fulfilling tasks. Eventually, yes, I think we'll delegate to AI in more and more complete ways, but it's a process that takes some time.
96.
by RA_Fisher 1y ago
We only need to do that when it’s practical for the task at hand. Some tasks are life-and-death, but many have much lower stakes.
97.
by RA_Fisher 1y ago
By moving up a level in the abstraction layer similar to moving from Assembly to C++ to Python (to LLM). There’s speed in delegation (and checking as beneficial).
98.
by RA_Fisher 1y ago
I chose it long ago and it was a tribute to the man. Then I learned about support for eugenics, and how he shilled for big tobacco by repeatedly insisting a causal relation between smoking and cancer was unidentifiable. He did certainly hav
99.
by RA_Fisher 1y ago
RTO seems like a bound-to-fail last gasp to enforce the patriarchy at work (and things that flow from it, sexism, racism, ableism, etc.).
100.
by RA_Fisher 1y ago
This claim's so big that it requires theoretical proof, empirical analysis isn't convincing (given the size of the claim). Causal inference experts have long known that many inputs map to outputs (that's why identification of
101.
by RA_Fisher 1y ago
For sure! Measuring parameters given data is central to statistics. It’s a way to concentrate information for practical use. Sufficient statistics are very interesting, bc once computed, they provably contain as much information as the data
102.
by RA_Fisher 1y ago
Yes, ChatGPT can do it using Python today (the statsmodels library). I use it all the time (I’m a statistician).
103.
by RA_Fisher 1y ago
LLMs can use Python to simulate from probability distributions. Though, admittedly they have to code and use their own MCMC samplers (and can’t yet utilize Stan and PyMC directly).
104.
by RA_Fisher 1y ago
> are we going to include the water used to grow the almonds you ate? That’s a good idea. Like nutrition facts but for everyday economic climate decisions.
105.
by RA_Fisher 1y ago
Competition is good and makes us all better off, don’t be afraid.
106.
by RA_Fisher 1y ago
Bc open source models aren’t as good and it’d cost a lot more than $100k to self-host given the same load.
107.
by RA_Fisher 1y ago
That’s what people said about SaaS. People pay good money for things they could theoretically do themselves with more time and money. Also, according to LLM Arena, all the best models are proprietary (bc trade secrets can be very powerful).
108.
by RA_Fisher 1y ago
> too far off the data distribution. I guess his prompts couldn’t provide sufficient information either (there’s no limit). Sounds more like a user issue to me. :) I don’t think there’s anyone that can type faster than ChatGPT.
109.
by RA_Fisher 1y ago
It’s a giant leap of technology? Investing in that is the perfect usage of debt.
110.
by RA_Fisher 1y ago
I'm a scientist and I'd never work that way. I'm methodical, because I've learned it's the fastest and highest-ROI approach. Guessing without understanding is extremely unlikely to produce the best results in a repe
111.
by RA_Fisher 1y ago
Why can’t they release it and let people decide for themselves?
112.
by RA_Fisher 1y ago
True, aside from the exceptional employers that’ll actually provide a good career.
113.
by RA_Fisher 1y ago
I like simplicity and I love a good base R idiom, but there's a lot less consistency in base R compared to the tidyverse (and that comes with a productivity penalty). Lattice is really low-level. It's like doing vis with matplotli
114.
by RA_Fisher 1y ago
Why? The tidyverse is so readable, elegant, compositional, functional and declarative. It allows me to produce a lot more and higher quality than I could without it. ggplot2 is the best visualization software hands down, and dplyr leverages
115.
by RA_Fisher 1y ago
Doesn’t the term have contractual obligations for OpenAI and Microsoft?
116.
by RA_Fisher 1y ago
He uses it instead of Google search.
117.
by RA_Fisher 1y ago
My dentist said he uses it near constantly, and so do I.
118.
by RA_Fisher 1y ago
Also, when asked the pilot said they didn’t switch the fuel control off. That seems like an important clue, bc it points at a possible glitch that could affect other aircraft.
119.
by RA_Fisher 2y ago
Another horrible cost of militarism.
120.
by RA_Fisher 2y ago
Location: Michigan, US Remote: Yes Willing to relocate: No Skills: extremely advanced Bayesian statistics / modeling and econometrics, causal inference (experimental and cases where exp are too costly) Resume / CV: 9th employee an

More ›