meztez

The best language for AI agents is the one YOU believe is the best.

Maybe some, there are also people who cheer for the capacity of humanity to figure out solutions that can benefits humanity. I view the perspective you describe as people externalizing their own insecurity. I do not view it as a critic towards people who have an excess of stored potential energy, but merely a signal that they haven't made peace with their own internal feelings about themselves. They need love, acceptance. They are on their own journey, everyone has the capacity to grow. Let's help everyone.

Seems complicated, although this is the best description of depression I've ever read, go outside and start walking.

https://docs.posit.co/, it just makes sense the way everything is organized.

A commercial food scale from Kilotech; quicker, more robust, always on, stainless, larger scale area and clear display.

For anyone who thinks about doing these, here are a few tips for the advanced CAS exams.

- Time is your enemy. - Use memorization techniques, like a memory palace. - Grader are looking for key words, be succinct. - If you do not know, skip. Your brain will figure it out while you work other questions. - Create your own study material from the syllabus. Anything that is remotely mathematical will be tested at some point. - Practice being efficient with the tool used to take the exam. Pen or keyboard.

Hated/Loved doing the exams to get my FCAS. It allowed me to explore my mental capacity. It teaches you how to learn fast, how to memorize, how to be quick at solving problems.

It has been a while, but after getting involved with the exam question writing/grading part, I felt the CAS was too entrenched in their processes. I stopped paying my dues since I did not plan to sign any acturial statement anyway.

I now do data engineering for other actuaries and maintain our tech stack. Think rating engine and ML in production. Code is mostly on github now, we have peer reviews, deploy to kubernetes. It is really is no different than SWE, except we know a lot more about the business side.

Life is great work wise.

5 years in. No regret so far. How long before you had regrets?

So this is the new bluetooth?

Fully agreed. Add `sessionInfo()` output as an appendix to your publication. Should not be too hard to rebuild from that.

R has data.table. It is the game changer as I agree base R data.frame do not cut it for performance. tibble will come close once they incorporate more of the data.table performance tricks.

https://h2oai.github.io/db-benchmark/

GPU integration was broken for a long time. Managing VMs / Environments. The absolutely horrible integration with git/Github.

Having to rebuild your environment from scratch when your workspace crashed. Imagine starting a notebook with a 45 minutes compile time. No go.

One click deploy, let's just forget about it.

It came down to IDE, workflow and data.table.

RStudio is an absolute killer solution from the get go. Package management in R is simple and robust. Shiny is the new Excel pivot table on performance enhancing code.

Python has more contributors, more users. It also creates a lot more noise. Business people may feel like it is a a programmer tool. R feel more approachable.

In the end, both are great solutions but we decided on R because we believe in the people contributing to the ecosystem, mostly RStudio. Somewhere down the line, there might be a transition to julia.

I'm pushing for our actuarial team to transition to more R + Git. After 3 years of preaching, most of the actuaries now use RStudio + git as their primary work tool. It is happening.

What we did :

1) Provide documentation on everything from install to using internal R libraries for ETL.

2) Provide mostly problem free, always updated VMs with RStudio Server/ Shiny Server.

3) Establish an hotline channel for instant help on R or git.

4) A couple members on the team developed really close working relationship with IT and we have great respect for each other work.

What we provide is way better and by being active, we built users trust in the tools.

We are phasing out SAS and proprietary modeling tools. Python never took hold even if we bought Anaconda entreprise. Excel is there to stay for sure but since actuarial student learn R in school, it is easier to onboard new hire.

If you want to go down this path and have a chat, hit me up. I'm in P&C. We use R both in development and production environments. We use it for pricing, spatial contractual obligation, claims assignment and a couple more models.

RStudio, server and connect are such a treat to interact with compared to anaconda entreprise.

A plumber in front of your ML serving function running in a docker container. There, simple, fast enough. We have multiple R containers in production, it is working fine and the data scientist are happy that they can keep working in RStudio.

Probably because I believe you should wear whatever makes you comfortable within the context of the event you are attending. Women could be wearing suits, and men dresses or skirts, I don't think that matters. And if and it was just down to my personal opinion, it would be wear whatever you want that makes you happy.

I understand some groups have expectation about people dressing nice and if I'm attending, I will conform.

But my point was that everyone had a dress code imposed upon them and that it could have been handled maybe a little better. Just tell people to dress up I guess, no need to specify which sexes gets to wear what.

And men told to wear blazer/sport coat, slacks, suit and 'step it up'. I don't see malice in the statement, it is just poorly worded.

It is interesting to see this discussion about notebooks while I'm thinking about all the RStudio users who do all their work inside the IDE and are pretty happy. Notebooks seem like such an inferior tool to me. I'm also extremely bias.

This comment has to be the most insightful I have ever read. It is true on so many levels and captures perfectly the mentality that some senior decision makers have about AI.

Were you using jsonlite? Great library that can parse directly from a URL and transform into data.frame depending on JSON representation. I agree it is not clear from the get go which library should be used when you do something outside of base R. You kind of need to be in the known

Believe me, I know. The code just becomes unreadable when you put all execution inside the same data step and use hash table to do fast small to big merging. And not to mention debugging that mess when you have a macro layer on top of it. Not having access to function source code, installation process being what it was. I do not miss it.

And yes technically SAS is faster than R but part of the equation is how many people can make SAS code faster than R/python. I had maybe, 1-2 people that could write efficient SAS code.

One version we had was a bunch of macro producing hash merge plus the whole how can I do something without having to get out of the data step. Just horrible. Number of characters in a line of code? You forgot your quote somewhere and now you have to run the magic line.

I hope I'm not too emotional when I say I hope SAS disappears from my industry and we embrace less adversarial licensing.

Debugging is something I could not do as easily in Julia vs debugonce and trace in R. Compiling takes time.

But so far we have seen great development. Flux is a truly beautiful ml library. Being a compiled language remove a lot of headache when building production images. The syntax, the full utf support in variable name. Package management is great. Having that abstraction layer between CPU, GPU so you don't have to rewrite code. Dispatch based on signature, type management. I don't see it going away soon. It took me 13 years to make them transition out of SAS, good thing cloud computing come around and someone realised the clusterfuck of having to manage SAS licence in the cloud.

One of the reason we use data.table is that it reduces the depencies when building custom images and its stability has been better than the tidyverse in the past. It might not be the case in the future, but that is how we made our choice initially.

Thanks, so far we just scaled up our vm ram but i might find a use for it.

I would highly recommend the use of the package data.table over tibble or the basic data.frame if you are doing any type of modeling in R with larger datasets. Yes R has many data structures but knowing how to use data.table will blow your mind in term of efficiency. Matt and other contributors have built something extremely fast and flexible.

I get that R is not for everyone but used correctly it is a beast.

Now this is anecdotal, but we have in the insurance industry what we call on level premium calculators. It is basically a program that will rerate all policies with the current set of rates.

Our current R program can rate 41000 policies a second fully vectorized on a user laptop that has a an i5 from 2015.

In contrast, the previous SAS program could do 231 policies a minute on xeon 64 core processor from 2017.

For our workload and type of work, R has been a godsend.

Bonus, we can put what our data scientist develop in R directly in production. (after peer review, testing, etc, not different than any other production code)

Back when I started in 2005, we modeled in some proprietary software like Emblem, used Excel to build a first draft premium calculator, rebuilt the computation in SAS for the onlevel program and sent specs to IT to rebuilt the program again for production. All three had to produce the same results.

I've tried Python, Go, Rust, Julia. I'd say Python could be a good alternative but speed of data.table, RStudio IDE and ease of package management in R makes R an obvious choice for us. I believe Julia to be the future but so far the adoption rate in house has been low.

R is like any other languages, we have a few rest API in production for live prediction. We use rocker docker image with xgboost and plumber, data.table to do pre prediction data wrangling. Hosted on GCP kubernetes, using 0.25 cpu and 250 mem, API is able to do around 40 requests per second per pod. Multi models, both have more than a 1000 trees.

I believe people underestimate what can be done in R/RStudio these days because they haven't been exposed to it.

R Markdown and its derivative are such great tools. I would also suggest xaringan if you want to present your work on a big screen.

Undeclared sales