BenoitP

43 ms·

451.
by BenoitP 5y ago
They do! But their lifespan is way shorter and their number is way higher; which means way more feedback from darwinistic evolution. To me it means their 'intelligent' behavior should be much more a function of being hardcoded in
452.
by BenoitP 5y ago
Hey, that's a rabbit hole I've gone through recently! I highly recommend the 'There Is No Antimemetics Division' short stories: http://www.scpwiki.com/antimemetics-division-hub
453.
by BenoitP 5y ago
If we're now a fertile ground for pandemics, it really does ask for a response in kind: changing another parameter in the differential equations. And that could come in shortening the response time. BioNTech had the vaccine weeks aft
454.
by BenoitP 5y ago
> if you were to not push the pedal at any of those instance That's the key to it IMHO. There has to be isolation for identifying causal relationships. Science is the work to be sure you're measuring the thinnest partial deriva
455.
by BenoitP 5y ago
I frequently have these retinal migraines; so does my sister. I'd say their drawing is quite accurate. In my personal experience: * their grayish color are way more yellow/white * it often forms a fully closed circle/potatoid
456.
by BenoitP 5y ago
> Aquabolt-XL: Samsung HBM2-PIM with in-memory processing for machine learning accelerators I wonder what kind of PIM: very simple row operations, like OR/AND/NANDing two vectors; or etching a close-by CPU in the same process l
457.
by BenoitP 5y ago
I think this is the equivalent to Flink's statefun: https://flink.apache.org/stateful-functions.html These are comparable to triggers on distributed tables under the hood, with a function addressing scheme and an ingre
458.
by BenoitP 5y ago
Taxing them would be the best of both worlds. Ultimately, proof-of-work is actually a proof of faith. How much resources are you ready to waste in order to prove your faith, and have a chance at discovering the next block? Now, a (globall
459.
by BenoitP 5y ago
Citus, cited at the end, is a column store (single node or distributed) https://github.com/citusdata/citus
460.
by BenoitP 5y ago
I don't know about linear algebra, but column major lets you compress thus: * Dictionary encoding: US,US,US,US,FR -> US:0,FR:1;0,0,0,0,1 * Run-length encoding: 0,0,0,0,1 -> 4x0,1x1 * Delta encoding: 0,1,2,3,4 -> 5x'+1&#x2
461.
by BenoitP 5y ago
Beyond the problem-solving and planning aspects, I think Factorio helps a lot in learning programming concepts. No theory, just experiments with visual and environmental feedback . * backpressure: your belts fill up * optimization problem
462.
by BenoitP 5y ago
One of the best retrofitting effort there is is Java's Loom IMHO. Instead of trying to solve the IO bottleneck on a few threads with async, it just reduces the Thread footprint and allows you to have millions of them. Good'ol Tomc
463.
by BenoitP 5y ago
> the unit allocates memory or not, the unit performs in a given worst-case time or not, etc. etc. -- but not everything it could do it also should Strong agree. This list could also be extended endlessly, and with continuous dimensions:
464.
by BenoitP 5y ago
I think streaming-first systems like Flink are best positioned to answer your need. You might need to abstract the query or subscription infrastructure into your data model though: Subscriptions: INSERT INTO user_orders_total_feed(use
465.
by BenoitP 5y ago
It produces vectors, also does the (distributed) searching. Here is a partial list of possible vector types: https://github.com/facebookresearch/faiss/wiki/Faiss-indexes
466.
by BenoitP 5y ago
Neat! I guess from the 'ivfflat' keyword that this project is serving vectors produced by FAISS[1] somehow? https://github.com/facebookresearch/faiss
467.
by BenoitP 5y ago
I've been thinking a bit since the last article from the author on Internal Consistency [1] on HN [2] on how to stress test a framework, or an application written on it. The problem is: it is hard to know in advance what kind of pressu
468.
by BenoitP 5y ago
Would that be the transform function from here [1]? I do agree too that this definitely should be a library on its own. Anyway thanks for your work; I'll definitely check SimpleR in my project. [1] https://github.com/ar
469.
by BenoitP 5y ago
Very nice! I've been prototyping something server-side in the style of the SQL Kappa architecture: facts on immutable logs, possibly keyed (and compactable by key), and SQL materialized views streaming the updates with Flink. And I&#x2
470.
by BenoitP 5y ago
> this would still leave the problem where each transactions causes two deletes and two inserts Indeed! You might be interested in the '6.5 Materialization Controls' section of this paper: https://arxiv.org/pdf&
471.
by BenoitP 5y ago
Very well written article. I'll add that there are tricks to manage that behavior. The dataflow is a diamond: transactions ----> credits \ \ ----> debits -----JOIN----> balance ---> tot
472.
by BenoitP 5y ago
> One even went so far as to state during an all-hands meeting with technical staff that ‘running a communication network’ was by no means a core competence for them. This is an outraging but very widely spread phenomenon. No industry is
473.
by BenoitP 5y ago
I'd love to see the React store and renders be animated through streaming SQL materialization. Think Flink or Spark or ksqlDB's views server-side, sliced and sent to tables client-side; where other views join these tables and anim
474.
by BenoitP 5y ago
Or about 0.04% of the market cap. Don't know about stock options, which must dwarf this gesture.
475.
by BenoitP 5y ago
Nice to see some new initiatives in this domain (or old ideas resurface), but there is a long way towards mainstream adoption IMHO: * My business clients and I speak SQL together. I don't see them learning a new language. I don't
476.
Rust async is colored, and that’s not a big deal(morestina.net)
2 points by BenoitP 6y ago|0 comments
477.
by BenoitP 6y ago
Dense vectors are pervasive in Neural Networks. Take a layer's activations and you have a new one (which is what NN embeddings are). Indexing them is really hard: The local neighborhood's volume to be searched grows by radius^dime
478.
by BenoitP 6y ago
> There's a 1:1 backing of crypto funds Well, this isn't difficult to prove. Do they have a public process to display that 1:1 backing? Google is failing me, I can't find anything but this: https://blog.coinbase
479.
by BenoitP 6y ago
Was looking for the same kind of information. Do they publish proof of ownership of underlying assets? Or are they going naked for some amounts? Do they consider themselves as a bank, and vie for a 8% non-naked rate? In the MTGox days, the
480.
by BenoitP 6y ago
Sort of. Here is the original proposal often referred to by Loom's architects: https://250bpm.com/blog:71/ Here is what the wiki has to say: https://wiki.openjdk.java.net/display/loom/Str

More ›