Something I've noticed especially with fable at work is it seems to be smarter, so it takes fewer tokens and burns less of my usage than smaller models would for similar tasks.
This isn't really on Google. Countries disagree on how things should be named. Google sources the US centric view from the Geographic Names Information System (GNIS).
So you prefer google to be the arbiter of what are official names within each country?
I think the US administration name change is childish, dumb and short-sighted, but if the authority of geographical names in an area designates it as such then I think it should show as such within maps within that area, at least as long as the clarification is readily shown. The same already happens for things like "Falkland Islands (Islas Malvinas)" or "Myanmar (Burma)".
I guess the question comes down to: do you want google to show the map that your current locality decides to be their interpretation or do you want google to be the arbiter of what a global real map is? Or is there a third alternative that I'm not aware of? (BTW the UN is not a solution to arbiter every countries and languages terms for every country/area/city/lake/ocean/etc.)
I wasn't asserting that, I was only pointing out that they can call any feature anything they like on their own maps.
But no, I don't want them to take a political stance on every naming dispute. In this case, however, when the "dispute" is between a clearly incompetent president and the nation he presides over, yes, seeing companies standing up against the madness would be welcome.
It was a very reasonable decision when it was made around 2008.
Do you really expect a US based company to go to legal war with the US government over something like this? Because there would 100% be repercussions it they did.
I think that's part of what is frustrating to so many people? Google isn't a small business, and has the legal resources to at least fight back if they were challenged by the administration. Any such challenge would likely fall apart in trial, but even if it didn't Google would have at least challenged the absurdity of it.
The person responsible for the GNIS corruption and the leadership of Google are in essence conjoined twins. The tech oligarchs are not going to make decisions based on interests that aren’t aligned directly with their personal enrichment; and the GNIS Corrupter-in-Chief wields such personal interests rather effectively.
It's wildly inconsistent. Surgeries, hospitalizations and specialists are mostly ridiculous. You can find affordable general/nurse practitioners if you know where to look and what to ask for.
Call around and ask "What is your self cash pay cost?".
Most people don't know to ask that and if you don't speak the magic phrase then you get billed at the rate the office would normally charge insurance. Not all places have self pay discounts.
Some states have special programs for hospitals but they vary wildly by state and you have to know how to navigate that particular program and be very patient. Some are need based so if you are above a pretty low income threshold you won't qualify.
Not exactly the same but I'm building a NAS box with old parts. Most of my spare RAM sticks are laptop DDR4 SODIMMs. There are SODIMM -> desktop DIMM adapters... it did not go well. The system would boot 1 out of 5 times. No adjustment of memory speed settings would make the system stable.
That's not surprising. At the speed DDR4 runs, PCB design has to account for all sorts of weird effects. Adding the additional traces on the adapter probably pushes all sorts of timing over acceptable thresholds. There's a reason one of the big textbooks in the area is subtitled "A Handbook of Black Magic". [1]
RAM is more like agricultural products (with short shelf-life) than commodities like fossil fuels, mineral ores, etc. You can manage an inventory or speculate on production, but you cannot really hold a "portfolio" of it in any sensible way.
So, you should get into RAM futures if you believe this is more than a transient arbitrage sort of situation. All extant RAM will become obsolete as the demand shifts to newer, fancier versions.
Yes but on what timescale? Replacing the > 5 year old sticks in my laptop would currently cost well over half what the machine ran me when it was brand new.
It might depend on the tech stack. I run a small niche website but it has PHP and a database (MediaWiki/PHPBB) and without Cloudflare I'd estimate I'd need to spend several hundred dollars a month to handle the traffic. Traffic used to be tens of thousands of requests a day. AI has increased that to between 400k and 3M requests per day but it's not a smooth distribution. This is with bot fight mode on that greatly reduces traffic.
I adopted Cloudflare because it was getting DDoSed by the AI crawlers. I'm pretty sure all of them are vibe coding their crawlers and don't bother adding rate limiting as a requirement.
That was my point. I was trying to be gentle by mentioning "unreasonable" things, but seriously — how did we get to the point where less than 6 requests per second (that's 500k requests per day) is considered a DDoS?
I've spent some effort on optimizing my sites, but most of the effort was focused on avoiding unreasonable (stupid) work. Do I need a session for every request? No, I don't! Do I need a database fetch for every access to my homepage? No, I don't! Is it a problem to actually load all of my static content in all supported languages (24) into memory and serve it from memory? No, it isn't!
I use Clojure behind nginx on the server for my sites. Oh, and I also pre-compress all static assets to Brotli, so anything that handles brotli gets a static file served directly from nginx. I also use immutable assets with unlimited caching semantics.
Really — the problem is that we've grown lax and our software has become bloated, slow, and with unreasonable code paths. If every page fetch does 12 database accesses and runs through a slow interpreter, that is surely going to be a problem.
That's the traffic after rate limiting controls and bot fight mode. It's 3-4 million requests per day without bot fight mode and just rate limits. And as I said it's not a smooth distribution. Plus the requests are almost never for pages in cache. It's always stuff like loading all the message threads from the year 2000 or loading up the details of every page edit ever made to a wiki page.
If it was more static content it'd be easier, it's really the db being a bottle neck in a dynamic site.
Yes, the software could be better optimized but then I'd have to own the development of it. There is no reason a niche website should be getting millions of requests per day.
Sure, depends on how accessibly to people you want it to be.
Most legit search engines are going to honor robots.txt and you can disallow access.
Next level would be using something like rate limiting controls and/or Cloudflare's bot fight mode to start blocking the bad bots. You start to annoy some people here.
Next would be putting the content behind some form of auth.
What is the positioning for this and how does it work? A comparison to SBE might be nice.
I understand the issue about using Layout and MemorySegment being verbose but the reason I'm using those things it to develop high performance software that uses off-help memory and bypasses object allocation.
What does "map Java record types onto native memory" actually mean? Did you somehow turn a Java record into a flyweight or is `Point point = points.get(0);` just instantiating a record instance using data read from off-help memory? If it's a dynamic mapping library using reflection, that's cool but doesn't it kill the performance goals for most Java off heap usage?
Is this more of a off-heap to heap bridge for pulling data into the normal Java space when performance isn't critical?
I use c-struct layout. I should be more explicit in the readme. I use classfile api to generate bytecode during initialisation of the Mem<T> and bytecode stored in cache in case if initialised again somewhere based on the same record type (I don't cache for records that are declared locally in a method). The class created from implementing Mem is a hidden class. So, basically, given a record, one can be able to analyse the layout based on record state description, and then for that Mem implementation (hidden class) we generate static final field varhandles + layout, segment is an instance field, and then generate bytecode the get and set to avoid reflection (actually, this is where most headache is in implementation). Go to the test package and see simple code for some adhoc rudimentary java (and native) files for benchmarks. Planning to test JMH benchmarks soon.
reply