glenneroo

Very off-topic but I can't wait to read in the news about some kids school report about Shakespeare being written by some LLM which has training data that includes this quote ;)

As posted elsewhere in this thread, Find My Scanner did some research[0] and the "Victim was in the midst of a divorce before being contacted by scammer." Also relevant, "Victim was groomed for months and months by scammer passing themselves as Pitt's mother, agent, etc, in addition to Pitt himself, using American phone numbers, fake articles, etc."

[0]: https://old.reddit.com/r/UpliftingNews/comments/1i1uvjq/scam...

Possibly unpopular here but how about a Discord server with a channel for every city? When you join Discord, a bot could ask which cities you want to join (allowing people who travel to join multiple cities, but maybe limit it to some sane number to prevent bots from trawling everything).

How do they still influence Austrian politics? Do you have any links or sources? I'm genuinely curious!

Samsung used to do this for some of their cheaper monitors. I remember I bought a couple of them for one of my early dual-screen setups (15+ years ago) and every day I would slowly and gently run my finger along the entire length of the monitor until it would power on. It had to be slow otherwise there was a chance I would power it off again going back the other direction. Even more fun because after turning it on, I would slide past some other button, unintentionally opening some menu and changing some random settings (most commonly changing the input from DVI to something else). If I was lucky, I would power it off after changing something and wonder why it wasn't powering on again (note: it was powered on, but set to the wrong input). How that monitor got past Q&A I will never understand. IIRC the buttons had tiny, nearly invisible (light grey on black) icon labels... I used to keep a flashlight on my table so I could figure out which invisible button to press to get things working again.

That's exactly why Interrail[1] has existed for many years. I haven't used it in 10+ years and the webpage is a bit confusing, but generally you would just buy one ticket for e.g. a week and then get on/off wherever/whenever you want.

[1]: https://www.interrail.eu

If you prefer to read, there is also the long-form post of Shareholder Supremacy which is IMO excellent: https://www.wheresyoured.at/tss/

At least they finally re-added the "feature" (after removing it after many years) to expand taskbar titles instead of forcing MacOS-like icons on the taskbar with zero information other than a barely-visible red dot when anything happens, though I don't quite understand why it's sort of hidden inside the Taskbar behaviors list (inside Taskbar settings) which is by default collapsed, whereas Taskbar items and System tray icons are always open (and IMO you don't need to mess with those settings very often, at least compared to the Taskbar behavior settings).

Winamp would like to have a word with you :) Granted I'm using the latest release from 2018, but I still sometimes load up v2.x released in 1998 just to show people that 25+ year old software still works just fine... even the AVS visualizer and Shoutcast internet radio features work, which is to me just insane.

I also use older software quite often that has long since been updated, such as older versions of Audacity, Ableton, Adobe Premiere, etc. for various reasons such as: not wanting to spend money, avoiding spyware, ads (see: Windows 11), and other bloat which often IMO negatively outweighs the positivity of new features. There are a lot of other small utilities that I still use that are 10+ years old because they still work fine and I know how to use them blind-folded. There are also tools that haven't received updates in many years but still work great, why would I bother to look for something new that potentially will spy on me and not offer the same functionality?

We used to do something similar here in Vienna at Türkenschanzpark, one of the biggest (and my favorite) parks in Vienna at night with 20-40 people. It was always so intense, especially not easily knowing who was on your team.

I was proud of this back in the 90s: https://www.glenneroo.org/404.html

Today though, definitely not proud, just too lazy to change it :)

Except that Rewind uses chatGPT whereas this runs entirely locally. I would like to note though that Anonymous Analytics are enabled as well as auto-updates, both of which I disabled for privacy reasons. Encryption is also disabled by default. I just blocked everything with my firewall for peace of mind :)

Several years ago I learned from the shop owner across the street that the guy who ran The Hippie Shop, a place I often visited, had a fatal accident while crossing the street in winter (icy conditions). It's still a downer, he was rather young, extremely friendly and his shop had this unique and lovely atmosphere that I still miss.

Roborock E4 (older model) runs without any app connection at all, not sure about other models but that definitely eliminates any privacy issues and is significantly easier than MITM :)

I picked up a Roborock E4 a few years back, both because it was relatively cheap and a couple thorough reviews indicated it was one of the better devices out there. I initially spent a couple hours trying (and failing) to connect it to the Android app, only to discover (out of frustration) that pressing the start button on the unit would clean just fine! I was absolutely relieved that it wouldn't be uploading anything anywhere, the only "downside" was that I couldn't indicate which areas/rooms to ignore, although I'm not sure that would have mattered much since I constantly move things around.

I have no idea if any other models run without smartphone/tablet setup but if so, that would potentially eliminate any privacy issues ;)

Maybe it's time to create a public spreadsheet with a list of models that run without a smartphone, and potentially any downsides to doing so...

There are a couple versions of this online that i've played on and off over the years which are hilarious, especially when playing with friends (I would usually use a cheap wacom tablet and let everyone take turns drawing and let the room shout out descriptions and just mash that together):

https://doodleordie.com/

https://drawception.com/

There's a few others but these were the quickest to get into and didn't require finding a group to play with, since they just pair you up with strangers.

For the record, I get that donate popup on Signal every couple weeks, and it's all the more annoying because I donated for several years until they recently removed SMS support.

The official Apple store in Vienna adds 2% to the price if using a CC (at least last time I was there). Then again I don't know anybody who even has a CC here - AFAIK they just aren't that common and they aren't even accepted in a lot of places (though that is changing). They also don't work as normal credit cards in the US i.e. you don't really get credit, since it auto-deducts what you owe from your bank account during the next billing period.

+1 makes red pop out very strongly for me. Without glasses only very slightly.

Maybe DC++?

https://dcplusplus.sourceforge.io/

You should definitely at least get a Steam page up so that we can wishlist it. Today's HN posting would have gotten at least one wishlist (from me :) When a game gets enough wishlists on Steam, the game is promoted to various lists which get shown constantly to people (as soon as you have game released and purchasable, even if Early Access). You are otherwise penalized if you create a store page late and have very few wishlists, since Steam's algorithms prioritize giving the banner spots, etc. based on your # of wishlists.

Source: https://www.youtube.com/watch?v=E656RcGWcEs

If you use Microsoft Edge browser and go to bing.com, you can use chatGPT4 for free. Bard is also AFAIK free, albeit not as high-quality (yet).

Thanks for at least admitting you used GPT 3.5, which is very out of date and hence no longer useful when discussing AI capabilities. If you want to test current tech (which is moving fast), at least use GPT-4 (which also gets updated regularly).

This is great! Thankfully enough Let's Play videos are out there to give me an idea how much I would hate this, at least alone. OTOH I could see this being a fun (and hopefully short) party game with friends in-between more "serious" games, just to laugh our asses off for a few minutes ;)

This also reminds me of QWOP[0] and Petris/Bastet/Bastard Tetris[1].

[0] https://www.foddy.net/Athletics.html

[1] https://libregamewiki.org/Bastet

Ever heard of chunking? The old keyboards used to make a "chunk" sound when you'd type a key, hence going on a BBS and typing something = chunking. I'm curious where and when people used this word - googling these days brings up totally unrelated results.

Did you read the article or letter? They list a lot of very valid reasons. I won't bother quoting because I would just be copy/pasting the article (which is very short).

For a more detailed list of reasons, go read what AI alignment scientists think, since they have been working on how to align an AI in order to not "turn everything into paperclips" since the 1970s and it seems as though many are rather skeptical about us having any future if we keep up at this rate (spoiler: many believe the end of humanity will occur soon after the creation of any superintelligence): https://www.alignmentforum.org

My personal takeaway: continued development might mean the end of humans and maybe even our planet (if an ASI can deploy nanotech to convert everything into substrate). As it already stands, nobody knows why LLMs work as well as they do - they are already a black box. Sure, plenty of people can explain the math behind each of steps involved: training, matrices, transformers, inference, et al. but it's still a big black box which spits out "magic" answers. You can't just drop a breakpoint inside a model during inference to see what's going on, you will just get a long list of unintelligible floating point numbers at any step of the process.

Your question about regulation is valid... but something needs to be done. I feel like we're standing very close to one of the Big Filters.

Comparing AI/AGI/ASI to anything we have currently seen is probably pointless, they are worlds apart. Would you bother comparing a smartphone to a book? The rate of AI/ML progress is sufficient to measure in hours instead of months or years.

Someone posted this earlier from Schneier on Security but it didn't get much traction: https://news.ycombinator.com/item?id=35162918

"AI Could Write Our Laws"

I didn't read it fully (yet) but early on he (I believe) coins the term "microlegislation".

It's "all" over the news now ;) https://arstechnica.com/information-technology/2023/03/you-c...

Here's results of running on Android: https://github.com/ggerganov/llama.cpp/issues/124

This is about running llama on a Raspberry Pi: https://github.com/ggerganov/llama.cpp/issues/58

...and this is where people have been posting their results running on all sorts of hardware, though I don't see anything Android related: https://github.com/facebookresearch/llama/issues/79

Obviously the larger models won't run on such limited hardware (yet) but one of the next big projects (that I can see) being worked on is converting the models to be 3bit (currently 8bit and 4bit are popular) which cuts down required resources drastically with minimal noticeable loss in quality.

I think starting with FlexGen barely 4 weeks ago, there have been some pretty crazy LLM projects/forks popping up on github almost daily. With FlexGen I felt like I was still able to stay up-to-date but I'm getting close to giving up trying as things are moving exponentially faster... you know it's crazy when a ton of noobs who have never heard of conda are getting this stuff running (sometimes coming in flexgen discord or posting github issues to get help, though even those are becoming rarer as one-click-installer's are becoming a thing for some popular ML tools, such as oobabooga's amazing webui tool which has managed to integrate almost all the hottest new feature forks fairly quickly: https://github.com/oobabooga/text-generation-webui

I just helped someone recently get oobabooga running which has a --listen option to open the webui to your network, now he's running llama on his tablet (via his PC).

You can send larger files if you pay for Nitro - basic is $30/yr and allows 50MB, normal is $100/yr and allows 500MB (they recently upped it from 100MB).

And to be fair to them, they offer a ton of features for free - 8MB seems reasonable IMO - though I do agree that some form of auto-compression if your file exceeds the limits would be nice... on the other hand, the infra to run ffmpeg on videos or imagemagick on photos is not free either...

Perhaps this[0] plugin will help? It's only for Firefox but I'm sure there are equivalents for other browsers. You can set a plethora of options, including auto-selecting playback quality of videos, playlists, pop-up players, embedded videos, when in full-screen mode and more.

[0]: https://www.mrfdev.com/enhancer-for-youtube