Product, Monetization, and CUDA – Arnon’s blog
-
Zero waste with SaaS products
… is there a way to stop customers using your product wrong? Earlier this year I sat down for dinner with Miia where we talked a bit about product management, and a lot more about zero waste at Ark, a zero-waste restaurant in Copenhagen. We naturally recorded ourselves getting progressively more drunk, and you can…
-
Running OKR workshops for dummies (part 1)
OKRs are a pain. Here’s how you’ll run your own OKR workshop, even if your team is remote.
-
5 things I learned while developing a billing system
Half a year ago, I joined Pleo, a FinTech startup with around 15,000 customers (we’re always hiring at Pleo!). I joined to take on and further develop the billing infrastructure. When I joined, I was actually a bit concerned about the narrowness of the new role and the new team. I mean, really… How much…
-
Top 6 must-read books for PMs getting into product management in 2021
Last time I wrote about books, I claimed I don’t read enough. Barely 7 books per year as of 2019. This year, that has changed. I more than doubled the amount of books I read this year, with 17 total books. In fact, 2020 has actually been pretty good for me, I’ll admit. At the…
-
Matching CUDA arch and CUDA gencode for various NVIDIA architectures
Updated July 12th 2024 tl;dr I’ve seen some confusion regarding NVIDIA’s nvcc sm flags and what they’re used for:When compiling with NVCC, the arch flag (‘-arch‘) specifies the name of the NVIDIA GPU architecture that the CUDA files will be compiled for.Gencodes (‘-gencode‘) allows for more PTX generations and can be repeated many times for…
-
No, your job isn’t to say “no”
A common trope I’ve seen repeated many times, is that it’s a product manager’s job to say no. This is wrong. It’s a very dangerous thing to get into young product managers’ heads. Saying no is absolutely not your job. You don’t need to say no. Saying no doesn’t accomplish anything. It doesn’t make you…
-
Using Tesla K80 cards, or AWS P2 instances? The end is near
A bit under the radar, NVIDIA’s CUDA 11 has dropped support for cards using sm_30 and sm_32. Furthermore, cards using sm_35, sm_37 and sm_50 have been deprecated, and their support will be removed in future versions of CUDA. These cards include the NVIDIA Tesla K40, Tesla K80, and Tesla M60 cards. Several of the companies…
-
2020 is here! Why are we still using 1980s tech?
Before COVID-19 hit in full force, I went to Warsaw for the Big Data Tech show to give a short 15 minute talk. For your viewing pleasure, here it is! Slides are available on SlideShare It's 2020 – Why are we still using 1980s technology? from Arnon Shimoni
-
Your DBMS is slow. What do you do?
I first used SQL for an online blogging platform I was building in Perl back in 1999. Moving from a flat-file to MySQL was truly a lifechanging experience. The (then) four-year-old DBMS had a nice little set of features and optimizations that I could easily apply and get a decent performance boost. Call me a…
-
My top 5 “product” books
I will admit I don’t read enough. I find it hard to focus on stories after spending my work-day reading things rapid-fire. Most of my day is spent reading industry reports, e-mails, Slack messages, product requirements, tweets, etc. However, ny recent shift towards being a better product manager has lead me to really dig my…