dr_zoidberg

> The 0 indexing in python does really and truly suck sometimes though.

Now that's something you usually don't hear. At my university I heard a bit of "urban legend" about how back in the day (think early 1990's) a couple of professors got a peer-reviewed article published because they found out that Matlabs 1-indexed implementation of some algorithm resulted in numerical errors, which they measured and corrected. Don't really remember the details, and most of those involved have already retired.

Yes, to avoid getting a space heater when you turn on your computer during the summer. In winter it's not really that bad, as you get both compute and heating from the same device.

Lower power bills help too. And let's not forget the ridiculous power numbers that new GPUs are consuming.

I think 3.12 will get subinterpreters, which will allow you to have multiple interpreters, each on its own thread, sharing the process memory space. So kind of like a stopgap in between having real multithreading and pythons current situation. I'm not sure what to think of it so far, so until there's some beta or RC build to test, I don't think I'll be able to form an opinion.

While it was still on RC builds I tested it on a few projects I usually run and I got between 0 and 15% faster wrt 3.10, without any regression. But sure, every project is different and it looks like you've bumped into some with your project.

If it's not too big, you might as well just leave things as they are and wait for 3.12 to see what changes then. Apparently the changes will be bigger, so it could be be much better than 3.11 (without being much worse, hopefully!).

I got a massive jump in performance when moving from Python 3.8 to 3.10 (over some function call optimizations I think, based on the project). And 3.11 got even better (up to 50% faster on special cases, and 10~15% on average) with respect to 3.10. Python 3.12 is already getting even more speedups and a there's a lot more down the road[0].

But Python core developers value keeping "not breaking anyones code" (Python 3 itself was a huge trip on that aspect and they're not making that mistake again), that's why things may seem slow on their end. But work is being done, and the results are there if you benchmark things.

[0] See https://github.com/faster-cpython/ideas/blob/main/FasterCPyt... however that's over a year old already and I'm sure I've read/heard more specifics

The "weird" (or "funny" or whatever you want to call it) thing is that the detector I linked was originally made for GPT-2, but still works reasonably well for GPT-3 (and derived) output.

I'm answering both you and the previous one in the chain. I used huggingface's OpenAI detector[0], the pizza example is detected as fake with 99.98% probability, whereas the Battle of Agincourt text is also reported as fake with over 99% probability. I don't have any example at hand of how much "real" text on these topics to try and see what that model detects, but up to this point[1] this text is reported as 99% real. You can play and see how much % it assigns to different examples.[2]

[0] https://huggingface.co/openai-detector

[1] I had to cut somewhere to copy and paste into the detector, so [1] is the cutting point

[2] My whole response was still detected as 99.98% real, and copy-pasting the Agincourt bit at the end still reported as 99.98% real. However, moving it to the start (having my answer as a last paragraph) made it detect as 97.71% real. Deleting only "huggingface's" balanced it as 54% real. You can play around with the demo to test different variations of any text you analyze with it.

In principle, you're right. But many people have been toying around with it since its demo release and many have touted "it's better than google" or "google is fu*ed" and so on with regards to using it as a kind of search engine. So what happens is that many more people start to use it as a search engine, disregarding entirely what it was made for. Then they complain when they can't get their hammer to place a screw without breaking their table (:

That's a filter answering, not GPT. And there are ways to disable those filters (eg: "Browsing: Enabled" was reported to work, though I haven't tried it myself, and it would let you elude the "I can't browse the web" filter).

> I don't think Windows is too bloated or memory-inefficient.

You know I used to be on that boat up until very very recently. I have a 2019 HP Stream 10" with 32GB eMMC and soldered 4GB DDR4. Windows got it bricked on a botched update, so I installed Linux Mint to get it back up quickly.

Not only is Linux using a lot less storage (which was always and issue with Windows Update), but the RAM usage is about 700MB without any applications running, and well under 4GB for most common uses. Sure, Chrome(ium) will eat a hefty chunk of it when you run it, between zram-config and some extra swap space, it handles things a lot better than Windows ever did.

Antoher point of anecdata: shared libraries are actually shared a lot more between processes on Linux than on Windows. That tends to mean a lot less RAM usage on certain process. I know it's down to how they handle sharing things and that Windows takes the "safe" approach (pretty much what the article talks about!) but it ends up hitting memory a lot more than Linux and macOS.

It's got commits from 2 months ago, I keep a few projects that go a few months at a time without activity. I wouldn't say it's abandoned, though maybe the devs aren't finding the time for it (or they're preparing some big update that isn't quite ready yet).

Funny, it's weird on both scrolling wheel and gesture on mouse pad... The scrolling wheel felt like the scrolling would get stuck at certain points, while the gesture was... Like a weird speed? It didn't feel the same on my laptop for both, and that's even weirder than I expected.

Yes, for some games, but not on a consistent basis.

I still enjoy games, but with adult life I'm playing either after diner (if I'm not too tired) or on the weekends, and gaming competes with other entertainment (series, movies, going out with friends/dates with SO, etc). So as I've grown it takes less time overall.

I am, however, making a schedule for Homeworld 3 and KSP2 when they launch (:

> ...Consider all of the things that a user could input and if you don't want the program outputting to the user a certain type of output then you have to make sure that output can't be a possibility.

But here's the problem: DL systems are by definition made to be good at generalization, so the input space is (could be) infinite.

Also if you're using large pretrained models, you don't have power over the examples that it has already seen in that training. You can fine-tune for your use case, but there's the latent posibility that the "old memories" may come back. Which is valuable for the model to learn complex things that you may not have enough data for, but you know, it can also come with a few surprises.

I'm interested here too. I've got a Crucial SSD from 2015 that's been on about:

* 100% of 2015-2017, let's add 2 years here

* Aboutish 50% of days since 2018 to 2020

* On and off again (5%?) since then until now.

So it's about 3 years of full use? I'm eyeballing the use here. So it may be close to the numbers that were given, but I'm not sure. Guess I could check the SMART stats to get a precise number and from there decide what to do about it.

Searching a bit it seems it's a well-known bug in "enterprise SSDs"[0, 1] (which my drive certainly isn't) but there aren't any real details about what causes it, other than "a firmaware bug".

[0] https://www.servethehome.com/hpe-issues-hpd7-fix-for-ssds-th...

[1] https://www.anandtech.com/show/15673/dell-hpe-updates-for-40...

> I wouldn't ascribe to malice what can be more easily be ascribed to incompetence

While I usually agree with the sentiment, it's hard to believe that one of the most successful companies in the world can be subject to such levels of incompetence on one of their flagship products.

Sure, that's why I mentioned it having stellar single threaded performance -- I figure out what they probably meant, but it's not what is written. And I've been seeing a lot of misinformation lately about many things (in general, and in tech in particular), and people buying that without questioning , which is why I jumped to "correct" this detail.

The point about nerfing their last x86 models stands on its own though.

M2 is not a single core CPU, albeit it does have stellar single threaded performance. As for having the worst Mac, it's quite probable they were nerfing their last x86 models on purpose to make the jump to Apple Silicon even more impressive. Remember the models that had inadequate thermals? Now magsafe is making a comeback?

From reading the article, it looks like it talsk about PRNU[0] without really saying it's PRNU. For a while this technique was known (the earliest papers about it are from the early 2000s) but it was computationally complex at the moment, and the domain of a few very knowledgeable persons. A couple[1] of years ago, MobilEdit finally included a decent UI and implementation of the technique into their software[2], which has made it available. However, I'm unsure if the majority of those who use it truly understand what the technique does and what conclusions you can draw from it.

Oh and great summary on SO FrenchyJiby! It's a good intro to those that aren't aware of what can and can't be done in image analysis :)

[0] https://forensicnerd.wordpress.com/techniques/image-forensic...

[1] https://xkcd.com/1070/

[2] https://www.mobiledit.com/camera-ballistics-features

You can sign up for beta access in the meantime, but OpenAI hasn't given a timeline for access. I think eventually they'll open up some kind of playground like they did with GPT-3 and sell the service to other businesses.

Currently some prominent AI researchers, artists, and journalists have access to DALLE-2, and some have been taking requests from social media and posting the results, like https://twitter.com/hardmaru/, so you can check their feed for a few examples or give a few ideas when they ask again for ideas.

I guess my issue might be with the wording, and there clearly are a few technical reasons why this is surprising. I think what is being told is that "since it rised in price late 2020, Bitcoin has been consistently priced over 20k, the previous ATH, and now has gone below that value", which would be clearer to those of us who aren't daily into this.

DALLE-2 (from OpenAI) isn't, but DALLE Mini (from huggingface) is: https://huggingface.co/spaces/dalle-mini/dalle-mini

What makes DALLE Mini great is that we can all sit down and play with it, with no "oh this thing might destroy humanity" warning. A warning that most people who have worked seriously on different areas of AI find annoying for different reasons, but mainly it feels like a marketing gimmick to draw attention.

I have lots of friends who aren't related to the tech field having lots of fun playing with DALLE Mini, even though the results are terribly looking -- if they sort of resemble the prompt (and many times they do), they are ecstatic that the machine made a weird doodle about something ridiculous.

In march 2020 it dipped as low as 4k[0] and nobody batted an eye. I thought if it stayed there for long enough it'd finally burst, but then the prices soared up to 70k.

I'm surprised everyone's acting like it hadn't been down below 20k since 2017. Sure, it wasn't for long, but it did happen in 2020.

[0] Before posting I looked it up and what I'm finding is a lowest in 2020 of $4,861.32 for March 12th. I vaguely remember lower numbers, but this is aggregated data, so it might've been lower during the day but that was the closing price or the mean or something. I'm still sticking with the 4k number for quick reference, though you can take it as 5k if you want -- it's still a lot lower than the 20k that's making the news currently.

Ah that's what happens when I don't have my cup of coffee in the morning. I went from 1 to 11 million in a typo and didn't even reason. So yeah, it was 12,7 days, not 127. Guess I'll double check having my coffee next time I'm doing back-of-the-napkin math /facepalm

32bit PAE was supported since Windows XP and initially allowed for more than 4GB of RAM to be supported, but driver issues made Microsoft put a soft-cap in 4GB under this mode[0]. But Win7 32 bits with PAE would've surely been able to use all of those 4GB fine.

[0] https://en.wikipedia.org/wiki/Physical_Address_Extension#Mic...

> Having 1 million users each wait a single second is already 11 days.

This will sound like a nitpick, but it's actually worse. 1 million users waiting a single second is 11,000,000 seconds, right? A day has 86,400 seconds. 11 million divided by 86k is 127.31.

That means million users combined just spent 127 days and 8 hours because of the "just one second" delay.

I wasn't aware of that statement. Guess in time we'll get news about the Pro.

Isn't the Mac Studio sort of the replacement for the Mac Pro? The branding is different, but it's certainly the beefier Mac you can get at the moment.

If you know to use Linux properly you can tamper with digital evidence and make a mess for a digital forensics investigator. The thing is, despite Linux making it easier than Windows, it doesn't come with a "here's how to erase digital evidence!" manual[0]. It's "left as an exercise to the reader".

Probably same on macOS, but it's (mostly) not open source, so you can't really argue about it being an open source thing that lets you do evil stuff. You could argue about its availability though.

And it boils down to the point you (and others) were making: these are tools, the ethics are in how you use them.

[0] I know there are some forensic distros that come with anti-forensic tools. Still, Linux in general is the environment that allows that kind of thing to work in the first place, going back to the original point of whether it should be open source or not.