usgroup

64 ms·

811.
by usgroup 8y ago
GitHub repo and a Makefile... Cut C++ binaries, statically linked, test on local against small data, and ship them to production with scp. Run either standalone, on GPU cluster or MPI cluster. How the data gets to where it needs to be depen
812.
by usgroup 8y ago
PII is radioactive. So: 1. Avoid it. 2. Delete it if you don’t need it anymore. 3. Keep it as coarse as your usage for it. 4. If it’s logs then pseudonomise and randomise. 5. If it’s sensitive content then Encrypt it per client and don’t ha
813.
by usgroup 8y ago
Anecdotal evidence doesn’t have to have any of those qualities but I do think that’s exactly right, yet most decisions just won’t have a basis in controlled trials due to the variation in the patient group and because you just can’t have co
814.
by usgroup 8y ago
I think individual dentists will often go on anecdotal evidence. A single dentist may fix 10k+ teeth or more and treat the same patients over 20 years or more and is therefore often well placed to consider what worked in their practice. Thi
815.
by usgroup 8y ago
... also remembering that a traditional dB will typically not store data raw. Row compression is normal and disk compression is normal . The typical column store advantage is block compression, predicate pushdown and column order storage.
816.
by usgroup 8y ago
Well SQL would present the ETL declaratively for one ... whether the output is denormalised or unconstrained has nothing to do with SQL.
817.
by usgroup 8y ago
The RDBMS advantage is that you can update your records and you can append to them without having to rewrite the dataset. That makes ETL much easier. Eg recalculate a column. It’s also that referential constraints can make sure your databas
818.
by usgroup 8y ago
I’d contend this personally. You can employ disk, or row compression on PG if you want. Compressed disk will actually make your queries faster. You can use cstore for ORC based column storage with PG if you want. Presumably the cost of a fe
819.
by usgroup 8y ago
Poker ... has dominant strategies. Has hidden state.
820.
by usgroup 8y ago
Dominant strategies in the game theoretic sense work regardless of what you do. Chess is a strategy game with no micro. So is anything turn based.
821.
by usgroup 8y ago
Well less than that for one and to do a whole lot more, at a far lower wattage :)
822.
by usgroup 8y ago
... what do you suppose will happen as OpenAI transitions to being a more commercially oriented company with a fiduciary duty to its investors?
823.
by usgroup 8y ago
“In total, the current version of OpenAI Five has consumed 800 petaflop/s-days and experienced about 45,000 years of Dota self-play over 10 realtime months...” Clearly very exciting and a remarkable achievement but does anyone else mar
824.
by usgroup 8y ago
Can anyone offer some context on “parapsychology” in general? Is there any credible results on this stuff that happened within the last decade? Etc.
825.
by usgroup 8y ago
TLDR anyone?
826.
by usgroup 8y ago
Well, you’re not maintaining the runtime environment ... if you’d like the minimum bytes for everything then consider writing your next app in assembler :) The purpose with multi language programming is articulating your solution the the le
827.
Facebook, Instagram, WhatsApp Outage(bbc.co.uk)
5 points by usgroup 8y ago|0 comments
828.
by usgroup 8y ago
JQ is a JSON specific query language . Not related to JavaScript.
829.
by usgroup 8y ago
2 general purpose languages: R and Python. Multi language solutions make sense with languages that specialise. So I'd never write in "Go and Java" for example, or "Clojure and C#" or some other pair of general purpo
830.
by usgroup 8y ago
That’s good or bad dependent on the world you want to see. In my opinion, engineers should have an ample toolbox for much the same reason that a good carpenter has a tool for any given thing. If well tooled engineers is what you’re after th
831.
by usgroup 8y ago
I think there's a bit of a threshold with language learning. When you know a handful (or just theo necessary ones for your job), there is probably a bit of tendency to be religious about your "stack". When you know a few more
832.
by usgroup 8y ago
Sure, and for any given application it’ll be possible to explain Markov chains as they apply to it. I recently did a financial valuation course where we did an “intuitive derivation of Itos formula” so that we could skip the measure theory
833.
by usgroup 8y ago
I think when you get out of the basic linear algebra and calculus prerequisite and into the analysis and measure theory prerequisite nothing takes a few months anymore :)
834.
by usgroup 8y ago
Unless you’re interested in continuous markov chains, infinite state spaces, renewal theory, excessive functions, and so on.
835.
by usgroup 8y ago
“No prior knowledge” and “explain all the rigorous maths” are mutually exclusive in my opinion. I stress this as honest advise to anyone reading. Rigorous maths is akin to trying to explain to your non technical friends what you do in devop
836.
by usgroup 8y ago
1. Elementary probability theory. 2. Poisson processes. 3. The Markov property. 4. Stochastic processes. 5. Realise that you’re missing a background in analysis, therefore you don’t know sh?t about measure theory but you actually need it to
837.
by usgroup 8y ago
Can anyone advise on whether it’s worth modelling general critical systems in TLA+ where they are not distributed? In broad strokes, how does one make sure that your real system indeed behaves like your model?
838.
by usgroup 8y ago
I think a useful definition of “elite” would be that you have some sort of competition the performance on which stack ranks the competitors, the top N of which are considered elite. Now, what that means in academia is pretty straightforward
839.
by usgroup 8y ago
Think of it this way, if this thing is real (and I doubt it is) then smart companies will make themselves the exception and buy hard into the 30+ market, and have much better employees as a result. 30+ women/men are the staple of scien
840.
by usgroup 8y ago
Sorry to hear about that. Here are some tips: 1. Bullies will keep punching you in the face until you fight back. You don’t have to win: just fight back. 2. You’re more likely to lose your job if you’re not respected and you’re less respect

More ›