At this point I'm pretty much trying to solve every problem I can using just S3. I'm currently writing code to manage client sessions by storing the invalidation set in a file on s3.
Working on Cloud SIEM at Datadog
Previously:
CEO/ Founder of Grapl, Inc.
Developer on InsightIDR at Rapid7
Security Engineer on DART at Dropbox
At this point I'm pretty much trying to solve every problem I can using just S3. I'm currently writing code to manage client sessions by storing the invalidation set in a file on s3.
Warpstream is really cool but their pricing just makes no sense right now. It's either "completely useless free tier" or "contact us for an enterprise contract".
@HalvarFlake Sorry to hear :\ best of luck to you.
„Firefox has enabled Cookie Banner Blocker by default in private windows for all users in Germany. Firefox will now auto-refuse cookies and dismiss annoying cookie banners for supported sites.“
ChatGPT being updated to April 2023 is huge for me. So many technologies I want to play with are ~1 year old.
I'm playing SSBM on Slippi and people literally quit when they see I'm maining Zelda lmao
This last guy was a DK, wtf is he quitting over.
So many "agile" processes ignore two critical components of the Agile Manifesto:
Individuals and interactions over processes and tools
Responding to change over following a plan
@sanjuanswan Wow, huge! Congrats on that milestone and I hope you enjoy the time off :)
This was a burn on Splunk btw
MGM Resorts claims ~100M will be lost due to the breach. And to think, they could have put that money towards a whole year of Splunk Cloud.
@castarco I think a major win is that I can share a rust script to someone and they can exec it directly. This is way better than playground if either:
A dependency is not in the playground
I can't share the code publicly for whatever reason
I feel like allocators so often have really solid documentation. snmalloc docs are so readable and cool and good.
https://github.com/microsoft/snmalloc/tree/main
Really interesting project in general.
@ekse I tried that but it didn't work, perhaps it was too short. I just tried again and it did work. Still, I find that unintuitive.
Thing that has greatly annoyed me today. Github comments saying "Commented Yesterday". Give me the time of day. There's a huge difference when a bot commented at 9am vs 12pm, I need that information.
bcachefs is looking for a new partner to fund their work
Sorting the strings lexicographically leads to about a 7% improvement but it's much more data dependent and it changes the performance characteristics enough that it's not really worth it.
When compressing (zstd lvl 14) some generated data I found that first sorting the data based on levenshtein distance led to a 10% size reduction. It's extremely slow though.
By comparison, sorting by the length of the strings led to a 5% reduction and was basically free.
It's possible that an approximate solution might be the right tradeoff but it's obvious that first sorting by length is a significant win.
@martijnarts Tons, and I don't think a lot really talk about it. Datadog has some nascent Rust codebases and I don't think we've ever really said so publicly - perhaps because it's best to avoid people applying just to write Rust? Dunno.
@funes Yeah, it's great. A few big companies are definitely investing, Microsoft seems to be one of the most serious ones. Still, I'd really like to see a company do more than just put devs onto Rust projects - like, hire people to work on the compiler, to build language features, to write docs, etc. These are things Mozilla did and I would definitely call those "the good times" in terms of the language.
TBH I think Rust desperately needs to have a "champion" company that centralizes some of the efforts and starts putting serious money towards dev.
Things have been rough since Mozilla dropped the ball in the hardest possible way.