Do the queries still have to be serial? I imagine so. The README is not clear but it can be interpreted as me being able to fire more than one query (say 2, but the first one is still running, like pg_background). Is that the case?
I've been asked this numerous times, and similar questions. Most of these questions boil down to "are you, the interviewee we are about to give a job to, a crazy person". Most likely you are not, so find some plausible scenarios, like a missed deadline, scope creep, or any of the normal shit that happens to us, humanize the other in the story, state the resolution (or what you changed in process), the end.
I don't know, you said isn't trivial then you said it's not trivial! :D It's a colossal effort if you are at Netflix size, they support probably the largest number of client devices on the planet.
> First, the fact that language models can be trained on large amounts of text data
and can generate human-like language without any explicit instruction on grammar or syntax suggests that language may not be as biologically determined as
Chomsky has claimed
"The fact that this advanced drone released in the year 2080 that can drive with the same agility as an eagle proves that the eagle's flying ability is not as biologically determined as some people claim.
In fact, any organism can fly if it sees enough data about flying!"
This is a new rephrasing of an older mantra, which is to use fewer words when possible. The writings of the greats in our field exude this quality, for instance those of Dennis Ritchie. Mind you, he does use the aformentioned words, but they are often necessary, for instance, using the wording "simple shell" is true, because the shells he developed were indeed simpler than the ones before, and provably so.
The only twist in this new take is that it has an issue with commanding language using such words, and I agree. If you are telling me how to do something, there is no need to qualify the effort, because you don't have that information and neither do I, so the language ends up being verbose and, more crucially, incorrect.
My favorite sentence on free will comes from Chomsky (who I think was probably quoting someone else): "if we don't have free will, then why are you arguing the point?"
It's also conceivable that companies running LLMs might consider various tweaks to popular questions as they appear in media so as to inflate the model's answers. No way to learn for sure, but competing companies where PR spells short term disasters, wouldn't put it past them to intervene when necessary. I am not familiar with how one augments and already trained model or if if there is even an architecture that's capable of that, short of input proxies in front of the model. I am curious to learn more if anyone knows.
Under some elementary scenarios (and using some Postgres offerings), you can do a full major upgrade using logical replication (even with replicas). It depends on your situation, generally I agree, any sophisticated use of Postgres (lot of views, materialized or not, fast changing schema, large objects, and so on) makes upgrades a pain and is generally somewhat of an unresolved problem in my opinion.
When one reviews another person's PR, it's as if you are reviewing your own. This is not a mere superficial verbiage; a few times, I'll review someone else's code, and actually block on small modifications and comment on them asking for more details, only to discover that the small change was to something I myself wrote and I forgot it so deeply as for it to appear entirely new. As one is kind to oneself often, one needs to be kind to the others, and trust them. It starts from there, and you co-learn and re-learn (which I think is a subtle point in the article), since whether you wrote it or them, it does not matter.
Oh yeah, it gets even more interesting in large refactors, say an involved rename that requires you check the context before you do so (like not just global search and replace). I do a search, this opens a buffer with all results, with file name, position, all that, and Emacs can allow you to "peek" into the file without even leaving the search buffer focus (i.e use a split screen). Then, use https://github.com/mhayashi1120/Emacs-wgrep to do the magic of in-place editing.
Golang runs some pretty serious software (docker, K8s, stuff at google, hashicorp, etc, all software that's functional and makes money). Rust is kinda the same, but not there yet, it has some benefits that are yet to materialize, but so is so much software out there. Does it matter? No, learning either will make you a better programmer. Learning either and NOT forming a strong opinion? you will be a saint and forever remembered.
At first I thought this was about how Netflix arranges movies for you to select, where each movie repeats down a horizontal line and you are forever scrolling right in cardinally equivalent but combinatorially different sets of movies. :D
He wrote his "Concrete Mathematics" book specifically to address your first point. I admit its the only book of his I read cover to cover and it wrecked me.
You are hung up on the wrong thing. Parent writes: "[clear consensus that] there was absolutely nothing unusual about it", I quote one of the greats saying the game is "more than impressive", which, at least we should agree, is the exact opposite of claiming there is "nothing unusual about this game". Onus is on parent to come up with a list of grandmasters claiming the game was "usual", at least I provided one refutation but there are many.
This is not true at all. For instance Nepominatchi commented on game saying Niemann's play was "more than impressive"[1]. Not commenting on the situation per se, merely your "clear consensus among other top chess players" comment.
Interesting. Transactionality is implemented via a regular thread lock, this means in a concurrent Rails app where this library is used in a hot path you might suffer some contention. Best is to use for marshaling data in non-hot paths such as stand alone scripts or app start up. I only say this because it's quite different from expectations around transactions in an SQL sense.
Now that you opened this forum for Spotify feedback: If I do "like/heart" a few songs and then go to the Radio based on one of them, please don't show the songs I already liked in that Radio. I mean, I already "liked/saved" them, why are they appearing in my discovery phase?
You used the word "GOAT", I gave you an example of another GOAT, showing that your argument that Carlsen is "not one of the goats" does not hold water because "he did not beat people convincingly" (itself quite a bizarre statement, the objective is to win, and Michael Jordan is one of GOATs because he won so many championships, and Karpov's argument can be made against him to, if you know the games played). This is not a straw-man, I don't think we have the same definition of that. Besides that one quote from Karpov, you cannot find anyone else making that point, so you used one quote for a a pretty substantial statement. If it was an opinion held by many, I might entertain it a bit, but it's just very wrong.
Both Jordan and Carlsen won championships. Many of Jordan's games were very tight, down to the last second. I can find another player who won more handsomely, that does not negate Jordan's achievements. That's the analogy.
What about RDS Proxy? I believe it's still only supports 13 which means if you use both the Proxy and Aurora you will have to wait? RDS Proxy does still mention only support for v13[1].
How do people square climate change and violent currents across EU rivers with Nuclear installations so close to such rivers (for instance the Blayais power plant)? I am genuinely curious. Is the idea that we will build stations that can withstand such forces? It's not just warming rivers, but also notoriously unpredictable conditions and our inability to make up an accurate model of what might happen. If someone has a study that takes this angle I am happy to read it.
I suppose tracing becomes more important in that new architecture, assuming of course that each service is logging the unique identifier for the trace (for instance the request ID or some system-initiated event ID), but of course that presupposes logging to begin with, so I am not sure what "incompatible with logging" means.