lmeyerov

71 ms·

901.
by lmeyerov 5y ago
Seeing the increased bandwidth is super exciting for a lot of business analytics cases we get into for IT/security/fraud/finance teams: imagine correlating across lots of event data from transactions, logs, ... . Every year,
902.
by lmeyerov 5y ago
Yes, most recently, any barnacle that can be stapled onto the monstrosity that is the Modern Data Stack. Before then, for Kubernetes. Before then, for the Hadoop ecosystem. (These all have value -- even we work here -- but the ecosystems ca
903.
by lmeyerov 5y ago
DO and friends wouldn't exist in the CPU side if that was true, so a spot economy in GPU side makes sense, and even more so bc they are higher-end time shares Even if we didn't have an open ticket with Azure since Christmas for mo
904.
by lmeyerov 5y ago
Our nodejs + python backends are all: multiple processes -> async/await nodejs/python event loops (used to be frp/streams) . It's good to solve event loop bottlenecks, but more from a individual query's latency p
905.
by lmeyerov 5y ago
RE:Windows, docker <> wsl2 ubuntu feels just like docker <> native ubuntu. We're able to do all our CUDA dev that way too, the only gap is OpenCL. I don't use docker desktop. As it's annoying to do most modern AI
906.
by lmeyerov 5y ago
If anyone wants to play with webgl/webgl2/webgpu/high-perf js, we're looking for a graphics engineer for all sorts of fun viz scaling projects (graphs used by teams tackling misinfo, global supply chains, money launderin
907.
by lmeyerov 5y ago
Google is coming from ~last place on enterprise+gov security relative to Microsoft and Amazon, which is maybe 75% of the market (and ignoring the Splunks of the world), and the ability to grow there requires real skills in services. Mandia
908.
by lmeyerov 5y ago
Yep. Some magic questions more at where enterprise pricing begins: - "Oh budgeting isn't until end of year.. what can we do until then?" Then you do a 50K thing for 5mo until thing under their personal or direct bosses author
909.
by lmeyerov 5y ago
We fled to the US from the USSR with nothing but the support of American + Italian volunteers and fellow immigrants for housing and job hunting. I'm sorry for Ukrainian and Russian citizens suffering from this - I have family from both
910.
by lmeyerov 5y ago
I'd have agreed with you 5-10 years ago. - I'm interacting with and hearing a bunch of regional tech companies, hospital networks, and worse, consultant shops, doing what you're saying. For the most part, they're not tha
911.
by lmeyerov 5y ago
Yes! To your point, I've been watching the slow uptake of basic ideas in science departments. In this case, neural networks are still largely looked at either with suspicion (wrong principles) or as magic (too difficult) by top non-C
912.
by lmeyerov 5y ago
Agreed. DeepMind's AlphaFold (graph neural nets that solved protein folding according to the community's own definition) is amazing... ...But as a former academic, it also a huge warning bell on the current effective capability ga
913.
by lmeyerov 5y ago
Official sources are less reliable: Russia is heavy on propaganda at the state and media levels, and destination country reports will be spiky, slow, or nonexistent. This presentation is interesting because it is a data presentation that is
914.
by lmeyerov 5y ago
Wrong thread? I use tools like that, but they're not the environments, credentials, & threat models (I think) we're talking about. I'd be curious if/how any data science teams (google/facebook/netflix&#x2F
915.
by lmeyerov 5y ago
Standard doesn't mean good, and the sophistication to do this well is beyond most orgs and individuals . You can come up with specific scenarios, but they'll get garbled through the telephone game over time. Much better to treat d
916.
by lmeyerov 5y ago
I like how an MIT CS professor put it to me once: He stays home to work, and goes to work to socialize. Most people are not digitally native, and as in my original comment, management layers are especially social. So I'd expect deep wo
917.
by lmeyerov 5y ago
Managers like it bc they don't do deep work, and do benefit from happenstance + zoom relief. A CEO manages managers, so even more extreme. (And I assume even more so in sales/marketing heavy orgs where folks aren't digital na
918.
by lmeyerov 5y ago
Graphistry | GPU visual graph AI for investigations | JS full-stack data viz, JS+WebGL perf, sales engineer, + more! | all levels | REMOTE (global/US) | full-time + contract | https://www.graphistry.com/careers --- Ant
919.
by lmeyerov 5y ago
Super cool! Funny enough... We've stacked up ~4 incarnations of how we back our different kinds of sticky live GPU session workloads, and ironically, a big one we aren't doing but I keep wanting to see solved is user-defined GPU c
920.
by lmeyerov 5y ago
That's graphistry's last 7+ years, where we do frontend CPU/GPU visual analytics experiences <> GPU server processes. 100% endorse this design, opens a lot :) Main tweak is our model grew to "small client browser G
921.
by lmeyerov 5y ago
At least in the US, the article discusses a valid but frustratingly incomplete step per-nation. There's a broken incentive system for doctors: they are in small/regional academic centers and get rewarded for recruiting, but mostly
922.
by lmeyerov 5y ago
If you are into webgl / js perf, we actually have an opening that would work well for this. We'd love full-time, but part-time for a long time is great too: build@graphistry.c*m . The project is encapsulated and multi-phase, so lo
923.
by lmeyerov 5y ago
One down round later and investor preference shares+rights means Tencent, and effectively China, owns. It sounds like revenue (profit) is nowhere near these levels, so long time for that to change. For all consequential purposes, their hu
924.
by lmeyerov 5y ago
Interesting, thank you! Why does 100M vectors not work in ES? - Is this a configuration issue -- common for ES users -- or something fundamental? - It sounds like latency is the main thing. Any numbers intuition here, and any other dimensio
925.
by lmeyerov 5y ago
I saw elastic has vector indexes now -- what does 'well' mean? My guess is for most popular uses, vector indexes are indeed more of a feature than a product, so will smooth out over time in big DBs, and curious what that should lo
926.
by lmeyerov 5y ago
- potential to push policy execution pushdown to the db, such as for authing bulk oriented queries vs record level - potential for use with RLS - potential for analysis tools In practice it seems boring, but in principle, interesting. I&#x2
927.
by lmeyerov 5y ago
Yes -- we actively work w/ DB/compute companies supporting Flight w/ ~ad-hoc APIs and are thinking through integration. But seen 0 with Flight SQL (and saw a ~20% lift with Turbodc vs ODBC). So curious about either Flight (ad
928.
by lmeyerov 5y ago
Exciting, remember whiteboard conversations about this years ago, and benchmarking early ideas like Turbodc :) I could not find: Is there a list of conformant DBs & their implementation statuses? Been curious both for Flight and Flight
929.
by lmeyerov 5y ago
We've passed on most of the google-pointing technologies in this thread precisely because of that architectural footgun Instead, we went with Casbin (microsoft research) because it can push to your DB (including multi-tenant-sharing if
930.
by lmeyerov 5y ago
Am curious esp wrt numba gpu jit...

More ›