Haven't seen anyone mention that project in a few years, but now I'm in the unique position to talk about it. I live somewhere where I can't get any proper internet service - mobile broadband is slow, DSL or fibre lines are not brought out to where I live, and the only other option is cable internet access, which I've 1. had bad experiences with in the past and 2. where I live is operated by a company with laughably bad reviews at exorbitant prices for what they offer. We are talking about 60 USD (eq) a month for 100 megabit service.

So I shopped around to see what other options there are, and Starlink made me an offer. Free equipment, which is usually 400 bucks, delivered to my house, and then an unlimited data plan at whatever speeds I can get where I live for 50 a month, with a one month free trial. I said yes, paid with Apple Pay (seriously, did not have to fill out a single form or sign anything) and the dish arrived the next day.

Now, I know, Starlink is run by Musk, who is somewhere around the top 10 of my nightmare blunt rotation and also pretty likely to be an actual neo-Nazi, but I say whatever. It's not like the alternatives are much better, and at least SpaceX has some actual value for humanity, if you ask me. I might put a "I bought this before Elon went crazy" on my router, though.

I got the dish delivered and set it up on my roof. The app - which is excellent - tells you to orient it north if you're on the northern hemisphere, and to roughly point it up. I built my own mounting solution - a wooden board with mounting holes that snaps in place on my roof - and set everything up, not expecting much.

I was absolutely blown away. The app, once more, is stellar and incredibly easy to use, and a joy to play around with. I got a satellite connection in minutes, and did a speed test. I got 200 down and 50 up in the Starlink app, but independent speed tests as well as my own experience routinely hit 400 down and around 80 up. Genuinely impressive. Ping around 30, by the way. Consistent as well.

The next few days were a similar experience, although I did notice a drop in speeds if there was heavy rain. The speeds dropped however to around 150 over 30, which is still more than usable, and latency was not impacted at all as far as I can tell.

Honestly, it's a super compelling package. Setup was so simple my grandma could have done it, the hardware is beautifully made and very robust, and the designers really did think of a lot here. The cables are just weatherproofed Ethernet and you can bring your own (although they don't recommend it), the router is Wifi 6 and looks damn snazzy, the dish can even heat itself up to melt snow in winter.

If you're looking for reliable internet service, I really can't recommend Starlink enough. If where you're planning on running it is within the service area and you're fine with the 50 dollar a month price point (no speed or data caps, by the way) I'd say go for it.

Now, there are people who will say that it's a good option for remote places, but not that great for densely populated areas in buildings that could get for example cable service, and you shouldn't rely on it. But, well, I haven't been completely honest here:

The real sting in the tale is that I live in one a large European city with plenty of access to other internet methods (just unlucky in terms of my specific building, which is getting fibre next year), and mounted the dish on top of my townhouse in one of the most dense districts in town. It works flawlessly, and it's been the fastest internet service I've ever had, period.

Course, it can't compete with a fibre line, sure, but many people don't have those - and then, service or hardware might still add large costs on top of that. And with Starlink, I can just take it with me whenever I move, and don't need to ever worry about ISPs again.

I don't have many sufficiently nerdy friends to talk about this with, so if you're curious or have any questions, I'll do my best to answer them. If you have Starlink too and feel like I missed something, feel free to contribute to the conversation.

Part of my role at work is in security policy & implementation. I can't figure this out so maybe someone will have some advice.

With the advent of AI coding, people who don't know how to code now start to use the AI to automate their work. This isn't new - previously they might use already other low code tools like Excel, UIPath, n8n, etc. but it still require learning the tools to use it. Now, anyone can "vibe coding" and get an output, which is fine for engineers who understand how the output should work and can design how it should be tested (edge cases, etc.)

I had a team come up with me that they managed to automate their work, which is good, but they did it with ChatGPT and the code works as they expected, but they doesn't fully understand how the code works and of course they're deploying this "to production" which means they're setting up an environment that supposed to be for internal tools, but use real customer data fed in from the production systems.

If you're an engineer, usually this violates a lot of policies - you should get the code peer reviewed by people who know what it does (incl. business context), the QA should test the code and think about edge cases and the best ways to test it and sign it off, the code should be developed & tested in non-production environment with fake data.

I can't think of a way non-engineers can do this - they cannot read code (and it get worse if you need two people in the same team to review each other) and if you're outsourcing it to AI, the AI company doesn't accept liability, nor you can retrain the AI from postmortems. The only way is to include lessons learned into the prompt, and I guess at some point it will become one long holy bible everyone has to paste into the limited context window. They are not trained to work on non-production data (if you ever try, usually they'll claim that the data doesn't match production - which I think because they aren't trained to design and test for edge cases). The only way to solve this directly is asking engineers to review them, but engineers aren't cheap and they're best doing something more important.

So far I think the best way to approach this problem is to think of it like Excel - the formulas are always safe to use - they don't send data to the internet, they don't create malware, etc. The worst think they can do is probably destroy that file or hangs your PC. And people don't know how to write VBA so they never do it. Now you have people copy pasting VBA code that they don't understand. The new AI workspace has to be done by building technical guardrails that the AI are limited to. I think it has to be done in some low-code tools that people using AI has to use (like say n8n). For example, blocks that do computation can be used, blocks that send data to the intranet/internet or run arbitrary code requires approval before use. And engineers can build safe blocks that can be used, such as sending messages to Slack that can only be used to send to corporate workspace only.

Does your work has adjusted policies for this AI epidemic? or other ideas that you wanted to share?

They all released videos at almost the exact same time, so even though I don’t care, I was made to care. Therefore, I’m inflicting that same pain on you. You’re welcome.

Mrwhosetheboss made a good point saying that the target audience for this thing are rich people who want phones that look flashy and can pay for them, but don’t care that they have worse specs than the less flashy ones at a similar price.

MKBHD called it the “S25 Ultra Lite”, which I thought was funny. He also brought up the issue of cooling. He additionally said that no one is asking for thin phones, although it seems that Apple has also bought into the idea that people want this, since it’s rumored that they want to release a thin iPhone this year.

Dave2D said that he tested the heat dissipation capacity of the phone and that it can handle itself well. Apparently it still somehow has a vapor chamber inside of it, as well as a wireless charger. Apparently it also has the smallest battery in the lineup, even though it doesn’t have the silicon carbon tech that is all the rage now. He made a good point though, namely that this could just be Samsung starting the trend so that the technology matures in a few years’ time.

All three of them mentioned that everyone uses cases these days, which immediately kills the whole purpose of buying a thin phone and losing out on better specs.

If you somehow have not gotten enough of tech YouTubers acting confused over Samsung launching a product that no one asked for, there’s also Techaltar and Tech Spurt. I recommend the latter for dirty British humor.

Hello everyone,

I wanted to provide a litmus test, of sorts. This test helps you figure out if you are engaging in the Sunken Cost Fallacy.

Sometimes I find myself asking if I should quit something, let something go, move on, etc.. It can be hard to figure out what the answer to those questions are. I heard a piece of advice regarding this very struggle, and wanted to share it with you all.

Ask yourself:

Knowing what I know now, would I still make the same decision that I did?

If the answer is "No" then you are most likely in a Sunken Cost Fallacy mindset. It could be argued that you should stop whatever it is you're considering stopping.

If the answer is "Yes" it is likely you are not in a Sunken Cost Fallacy and you made a decision in line with your values, even if it may not seem like it (hence the struggling).

Disclaimers:
Knowing the answer to this question does not imply you will know what to do with the information.

This test does not really "fix" anything, so to speak, but it is intended to help you realize when you are in a sunken cost fallacy situation.

Message for Deimos or anyone else on Tildes development

I'm putting this here after writing the rest of my post, but could we please get an "expand edit window" option, unless there is one and I'm blind? The preview window's great but the edit window's locked to 6 lines and I would appreciate some more since it would make editing a long post like this much easier. Thanks.

The actual post

There has been a lot of gaming buzz regarding extraction shooters as of late, with the closed alpha of Bungie's Marathon currently underway, the second tech test of Embark Studio's ARC Raiders starting just now, and the recent news of the cancellation of a Titanfall extraction shooter from EA/Respawn. As someone who's played and enjoyed extraction shooters before I've been following these and observing the discourse surrounding them (except the Titanfall one, I didn't even know that was a thing til the news of its cancellation) and I've been somewhat dismayed to see a lot of confusion, mixed messaging, and general disdain for the genre. So I've decided to put my own thoughts and definitions of it down here, and clear up the most common misconceptions or falsehoods I've seen repeated ad nauseam elsewhere.

A brief introduction to extraction shooters

First off, what IS an extraction shooter, what makes it different from other shooters, notably battle royales, and which games qualify as extraction shooters? An extraction shooter, as its core gameplay loop, is a shooter where you enter a map with loot and AI enemies scattered about, and the goal is to gather loot and extract from the map with it. However, you need to get out alive - should you die, you will lose everything (with some exceptions) in your inventory, including the gear you went in with. On top of that, the most popular and successful extraction shooters are PvEvP - you will be competing with other real players for loot, and taking loot off their bodies can be just as profitable if not more so than taking it from PvE enemies. It is optional though, and it is entirely viable to play as a "rat", sneaking around and gathering loot without drawing attention and extracting without anyone noticing.

Not a battle royale

Extraction shooters are also frequently confused with battle royales as both games have players inserted into a PvP map where they scavenge gear. However, the similarities more or less end there. With battle royales, you do not risk losing your items on death as all players are dropped in with nothing and don't have a stash to draw from or store items in, so any "loot" found is merely a means of securing victory for the current round. PvP is also mandatory, as the goal is not to get loot but to be the last team/person standing. To facilitate this in a timely manner, battle royales have a shrinking map mechanic that forces the remaining players into a smaller playable area as time goes on to force them into a confrontation. Extraction shooters do not force PvP or have shrinking maps but do have their own ways of drawing players towards each other, through loot-rich points of interest and extraction zones. Some parts of the map will have greater quality and/or quantity of loot, which will naturally draw players in, and there are a few designated areas where you can actually leave with your loot which will also increase your odds of encountering other players either trying to take your loot before you can leave, or trying to leave themselves. Because it is not forced though, PvP encounters are a much more unpredictable and organic experience in extraction shooters.

What extraction shooters are out there?

So which games count as extraction shooters? The current leaders in the genre, which also happen to be some of the longest-lasting ones, are Escape From Tarkov (EFT) and Hunt: Showdown. EFT is a rather hardcore modern military FPS with a heavy focus on realism - guns are extremely customizable, ammo types and armor can make a huge difference, bullets are extremely lethal even from AI enemies, and a good headshot will drop even the most geared and armored player so there's always risk. It has a cult following but its hardcore emphasis makes it unapproachable for most. It also has periodic progression wipes where players have to start over from scratch to keep things fresh and more fair for newcomers, but is a major turnoff for players that don't like to lose what they've earned. Hunt is an FPS set at the end of the 19th century with a bit of dark magic/voodoo theme. Guns are reflective of the times and rather limited in terms of rate of fire and reload speed, which results in more drawn-out firefights where every bullet counts. For each round, the focus isn't to get loot around the map but rather to track and hunt down a bounty boss monster, then extract with that bounty. These two games are what will come to mind first when extraction shooters are mentioned, EFT more so.

I won't go over cancelled (Titanfall), discontinued (The Cycle: Frontier), or side game-mode (The Division's dark zone/survival) extraction shooters here, which is basically almost all of them sadly, so I'll talk about the two biggest up-and-coming ones instead, Marathon and ARC Raiders.

Marathon and the surrounding controversy

Marathon is a sci-fi FPS that uses the lore of Bungie's Marathon trilogy from the 90's as its setting. You play as a "runner" in a robotic shell scavenging the remains of the colony on Tau Ceti IV for scrap to fulfill contracts for the megacorporations involved in the colony's development who now seek to find out what went wrong. It checks the usual boxes for an extraction shooter - you go in with your own loadout, scavenge at points of interest to fill your limited inventory, defeat PvE enemies and other runners for loot opportunities, and try to extract alive before time's up. There are a couple things of note that have resulted in mixed opinions:

• The art direction for runners, gear, and architecture is a sort of mass-produced, neon-colored, smoothed plastic, blocky style, which is a "love it or hate it" kind of thing.
• The gunplay is very similar to that of Destiny, Bungie's last game, which in my opinion is very solid. However, they did make the decision to have mouse magnetism enabled for PC (your cursor will magnetize to targets) to give it more parity with console players, and that has been very unpopular.
• The only queueing option is teams of 3 and the devs have stood their ground on not having a solo or duo player queue, which is a turn-off for players that prefer solo, or don't want to play with randoms and don't have 2 friends to play with.
• Players do not have full customization of their runners but must choose from 6 runner archetypes that have a set of abilities and a specific look, which can be partially changed with a skin. This is also largely unpopular, as Bungie's past titles have featured high levels of cosmetic customization and this feels largely restrictive and monetization-focused instead.
• There is no option for proximity voice chat, which the devs have said was excluded to protect players from toxicity. This has also proved to be very unpopular.
• It will have seasonal progression wipes which will reset faction reputation and clean out player vaults, which is unpopular among players that like to hoard their loot, especially many Destiny players.
• Supposedly there will be more "raid-like" PvE experiences on an as-of-yet unreleased map that takes place aboard the Marathon colony ship, but how mechanically complex those are or whether or not that will be enough to attract PvE-oriented players is pure speculation at this point.
• It will not be free-to-play, but rather released at a "premium" but not full game price point, which most people are assuming to be $40 USD.

The game is set to release in September this year, but based on the feedback Bungie is getting from players in this very first alpha, they will need to take this feedback very seriously and make a number of changes in the few months they have left, or risk a very rocky release and potentially financial failure. Many players seem to want Marathon and Bungie to fail, notably vitriolic Destiny veterans that feel like they were snubbed out of Destiny 3 for this, but as someone with over 2000 hours in Destiny 2 myself I want it to succeed, whether I play it or not. I'd rather there be more fun and successful games than major failures, and wishing for something to fail just because it isn't what you want is incredibly petty.

ARC Raiders, the underdog

ARC Raiders is a third person shooter set in a post-apocalypse where robots called ARC have devastated the surface of Earth and humanity has retreated underground, sending "raiders" to the surface to scavenge for tech and goods. It's developed by Embark Studios, which is made up of ex-DICE (Battlefield) developers, and their other title is the well-received but niche PvP shooter The Finals. Mechanics-wise, there isn't anything particularly unique about this extraction shooter - limited mobility, limited inventory space, PvE enemies, points of interest, extraction points, etc. However, it seems to check all the boxes of what players want and it does it well while making the experience more casual and accessible:

• There are "safe pockets" where players can store a few loot items they won't lose on death (Tarkov does also have this, Marathon and Hunt do not).
• There is proximity voice, and also a quick emote menu for giving vocal commands, directions, and responses.
• The art direction is realistic post-apocalypse with high graphical fidelity and semi-futuristic robots, which is "safe", but still good looking.
• The audio design is phenomenal. Distant gunfire, supersonic cracks of rounds, bullets ricocheting off surfaces, large bots stomping around from blocks away, player footsteps on different surfaces and within enclosed spaces, quadcopter drones buzzing, larger bots with unsettling and deep "roars", and more.
• There isn't a solo or duo-only queue, but there is solo-matching priority at least, so you're more likely to be placed with other solo players.
• There is a clear progression path in the form of a workshop you can upgrade with scavenged materials, and a deep skill tree with multiple branches.
• Raiders are not class or hero-based and are freely customizable by the players in terms of loadout and appearance.
• Chaff PvE enemies are relatively easy to defeat but still dangerous. Aside from eliminating them as threats, they can drop materials or items to recharge your shields so they're not just a waste of ammo. Non-chaff PvE enemies can be significantly more dangerous and harder to defeat but will have valuable loot.
• There is already an example of a PvE "boss" robot guarding a point of interest which requires some mechanics to collect the loot within it. The boss itself is very challenging to defeat, even with multiple teams fighting it, but should reward top-tier loot.
• While the game was initially announced as free-to-play, the devs have since switched to selling it at at $40 USD.
• There are "battle passes" in the game in the form of shops players can unlock using a currency that is earnable in-game, or purchased with real money (like warbonds from Helldivers 2 for any helldivers reading this).

Overall, the game is shaping up to be a more accessible extraction shooter for the wider gaming audience and very serious competition for Marathon. No official release date has been announced but they are planning on releasing some time this year.

The stigma around extraction shooters

I've mentioned various things about extraction shooters that may be contributing to their unpopularity amongst the wider gaming audience throughout the post, but for the sake of cohesiveness and for all the folks that just want a TL;DR, I will collate and expand on those ideas here:

• Confusion with battle royales - I've seen some people confuse extraction shooters with battle royales and say "the market is oversaturated with extraction shooters, dead trend chasing game" or something along those lines. It's hard to call a market oversaturated when there's only 2 successful and very niche games in it, but if you incorrectly lump all the battle royales in that makes more sense.

• Escape From Tarkov is a bad poster child - When people do think of extraction shooters (and not mistakenly battle royales), they will default to EFT, which is notoriously hardcore and "sweaty". It would be the same as never having played an RPG, and being introduced to it with Dark Souls, which would understandably turn away anyone that isn't looking for that kind of experience.

• PvP and losing progress - The game always having a PvP element is already discouraging to PvE-focused players, and this is only made worse by the chance to lose your gear if you die. Many players are strongly opposed to losing progress, and losing multiple times in a row due to other players defeating them when they just want to do PvE and get some loot is an awful experience that they don't have to have in a different type of game.

• Progression wipes are anathema to hoarders - On top of potentially losing progress on a round-to-round basis, seasonal progression wipes also threaten to reset progress entirely between seasons, While they are effective at keeping things fresh, players that like to have 400+ items stored away in their vault that they will never use and just admire from time to time revile this concept. Less hoard-minded players may be concerned about their potential inability to max out their progression, the fear of missing out induced by temporary progression, or the pointlessness of even progressing if it gets wiped anyway.

I didn't read any of that, are extraction shooters for me or not?

Well, that depends. If you:

• Don't mind or enjoy a mixed PvEvP experience
• Are ok with the idea of loot and progression being temporary
• Like the higher stakes of potentially losing your inventory and gear, or taking someone else's
• Are ok with inventory management, even in the middle of a round
• Are not terrible at shooters

Then yes, extraction shooters may be fun for you! They certainly aren't for everyone, and there's nothing wrong with not enjoying one or the genre in general, but if you do, they offer a very unique gaming experience. If you are interested, keep an eye on ARC Raiders and Marathon - they aim to be more accessible than previous extraction shooters and it's a lot easier to get in on a new game than join one with a veteran playerbase.

Question is as stated in the title. How did you learn to read?

I am re-listening to the great podcast, "Sold a Story" and it has prompted a lot of questions to myself, and now to others.

So, I'm curious, how did you learn to read and what do you remember about it? I am extra interested in people who have learned from "non-phonetic" languages, and also have a new curiousity about French, which I consider a language that does not match the spelling of its spoken and written words (if that makes sense, I'm sure that is my own bias there, as an English speaker).

I dont think it would be unusual to say that I enjoy life when things are running smoothly and everything feels under control, stress levels are low and I can plan for an enjoyable future without much worry.

And then everything goes to hell in a hand basket. Like being wracked with unimaginable pain so bad I wake my wife in a cold sweat at 2 am and choke out "We need to get to Emergency now". And then, unbelievably, it gets even worse, so bad that thoughts that death might actually be sweet relief start to creep in.

That was two weeks ago when I found out that not only did I have a 3 cm gallstone stuffing up my gall bladder but it had perforated into my liver and my gut was filling with infection, a condition that can shut down organs or even be fatal if not treated rapidly. Through the miracle of modern Canadian healthcare, they had me multi tested, diagnosed and into emergency surgery in short order.

And a fortnight hence, I have a lot to ponder (because Im still too damn weak to do much more than type) and its made me admit that unexpected adversity is a gift not a curse.

Foremost, it focuses the mind. When youre laying on an operating table surrounded by surgeons and nurses and wondering if you're going to come out of it alive, a lot of things become unimportant. I didn't care about politics. Or bills. Or investments. Or achievements. Or just about anything. I just wanted to be ok, not only for my own sake but especially for those I care about. And at that point there was crystal clear realization that what counts is only that - those I care about. The rest is dust and meaningless in the grand scheme of things. A lesson I've learned profoundly once before, but the mind dulls with an easy existence and needs a refresher on occasion, unwelcome as it may be.

Coming through also taught me how much I take for granted, especially having reasonably good health. I've had random unexplained attacks before, but for a day I dealt with incredible pain and it was unbearable. I had to think of the people who deal with that kind of soul crushing challenge continuously - their existence and will to persevere is challenged on a daily basis. And hardly anyone sees that exhausting internal grind but just having the will to stay alive is a hard won battle every single day and no one's handing them trophies for it. I have respect for those who do it, and a much greater understanding for those who just can't and decide to opt out. I get why that makes sense for some.

I also have a newfound debt of gratitude to people with character, foresight and undefeated willpower like Tommy Douglas who fought for universal healthcare in this country, against the will of most doctors at the time who (to my great surprise) actually went on strike to oppose him. After 20 tests, xrays, a CT scan and emergency surgery (with 2 surgeons, anesthesiologist, and 4 nurses), and multiple days recovery in big, brand new private room and being sent home with all my meds my entire bill was zero. No one even mentioned money and there is no insurance or co-pay to settle. Its done. I cant imagine the burden Id be feeling today if I was now saddled with crushing debt, but I am deeply grateful for the system that did all this for free, even if I do have to pay higher taxes to get it. I will remember that the next time my income tax bill comes around and make a mental note that my taxes are not 'wasted'.

I'm not going to be yodelling with joy if something this painful slaps me upside the head again anytime soon. But I also meekly acknowledge that sometimes life's most profound, most well remembered lessons dont come out of joy, they are often seared into memory by unexpected, even shocking adversity. I might not like it at the time, but in hindsight, it's a gift. An unwanted but valuable gift.

There are two relevant articles, both of which provide insight into why these changes were made.; they are relatively lengthy, but I have provided summaries of the changes.

Commander Banned and Restricted Announcement – April 22, 2025

Summary: 5 cards have been unbanned and are now on the game changers list

• Gifts Ungiven
• Sway of the Stars
• Braids, Cabal Minion
• Coalition Victory
• Panoptic Mirror

Commander Brackets Beta Update – April 22, 2025

Summary: no major changes to the mechanics of how brackets function. Two cards have been removed from the game changers list and 18 cards have been added.

• Trouble in Pairs removed
• Trinisphere removed

The following cards have been added to the game changer list: Teferi's Protection, Humility, Narset, Parter of Veils, Intuition, Consecrated Sphinx, Necropotence, Orcish Bowmasters, Notion Thief, Deflecting Swat, Gamble, Worldly Tutor, Crop Rotation, Seedborn Muse, Natural Order, Food Chain, Aura Shards, Field of the Dead, Mishra's Workshop

Examples:

• A live service game that got shut down
• A multiplayer game that no longer has a live community
• A game whose cultural relevance has faded
• An older game that doesn't stand the test of time
• A game you had spoiled for you
• A game that got updated and went in a different direction
• etc.

Let us know what it is, and why you feel that you missed out on it.

The Issue

I'm streaming games from a desktop PC hardwired into my router (running Sunshine) to a laptop wirelessly (using Moonlight). It works beautifully.

Except, every five minutes, the stream chugs: my framerate drops precipitously, and Moonlight gives me a warning telling me I should lower my bitrate. This happens for only a few seconds, before it resolves and goes back to normal.

I timed the interval between the chugs several times and got approximately 5:07 between each slowdown. It is remarkably consistent.

Because it's so consistent, I assume there's some scheduled task or something running every five minutes that's causing it to chug. Dropping the bitrate makes the chugging less noticeable, but it still happens.

Ruling Things Out

I think it's safe to rule out the idea that it's my router or the host PC.

I have a smaller 13" laptop that I used to stream to, and I just recently bought a 17" to replace it. The five-minute issue only happens on the 17", even with identical stream settings (same resolution, FPS, and bitrate).

The computers are obviously different hardware, but they're also running two different linux distros.

The 13" Laptop is running MX Linux 23.5 (KDE). This is the one that works.

System:
  Kernel: 6.1.0-32-amd64 arch: x86_64 bits: 64 compiler: gcc v: 12.2.0
    Desktop: KDE Plasma v: 5.27.5 Distro: MX-23.5_KDE_x64 Libretto September 15
    2024 base: Debian GNU/Linux 12 (bookworm)
Machine:
  Type: Laptop System: Dell product: Latitude 7370 v: N/A
    serial: <superuser required>
  Mobo: Dell model: 0XFY7T v: A00 serial: <superuser required> UEFI: Dell
    v: 1.28.3 date: 02/07/2022
Battery:
  ID-1: BAT0 charge: 12.6 Wh (62.1%) condition: 20.3/34.0 Wh (59.6%)
    volts: 8.1 min: 7.6 model: SMP DELL WY7CG58 status: charging
CPU:
  Info: dual core model: Intel Core m5-6Y57 bits: 64 type: MT MCP
    arch: Skylake rev: 3 cache: L1: 128 KiB L2: 512 KiB L3: 4 MiB
  Speed (MHz): avg: 2496 high: 2758 min/max: 400/2800 cores: 1: 2400 2: 2758
    3: 2400 4: 2429 bogomips: 11999
  Flags: avx avx2 ht lm nx pae sse sse2 sse3 sse4_1 sse4_2 ssse3 vmx
Graphics:
  Device-1: Intel HD Graphics 515 vendor: Dell driver: i915 v: kernel
    arch: Gen-9 bus-ID: 00:02.0
  Device-2: Realtek Integrated_Webcam_HD type: USB driver: uvcvideo
    bus-ID: 1-9:5
  Display: x11 server: X.Org v: 1.21.1.7 with: Xwayland v: 22.1.9 driver: X:
    loaded: modesetting unloaded: fbdev,vesa dri: iris gpu: i915
    resolution: 1920x1080~60Hz
  API: OpenGL v: 4.6 Mesa 24.2.8-1mx23ahs renderer: Mesa Intel HD Graphics
    515 (SKL GT2) direct-render: Yes
Audio:
  Device-1: Intel Sunrise Point-LP HD Audio vendor: Dell driver: snd_hda_intel
    v: kernel bus-ID: 00:1f.3
  API: ALSA v: k6.1.0-32-amd64 status: kernel-api
  Server-1: PipeWire v: 1.0.0 status: active
Network:
  Device-1: Intel Wireless 8260 driver: iwlwifi v: kernel bus-ID: 6c:00.0
  IF: wlan0 state: up mac: <filter>
Bluetooth:
  Device-1: Intel Bluetooth wireless interface type: USB driver: btusb v: 0.8
    bus-ID: 1-2:2
  Report: hciconfig ID: hci0 rfk-id: 1 state: up address: <filter> bt-v: 2.1
    lmp-v: 4.2
RAID:
  Hardware-1: Intel 82801 Mobile SATA Controller [RAID mode] driver: ahci
    v: 3.0 bus-ID: 00:17.0
Drives:
  Local Storage: total: 238.47 GiB used: 31.99 GiB (13.4%)
  ID-1: /dev/sda vendor: Toshiba model: KSG60ZMV256G M.2 2280 256GB
    size: 238.47 GiB
Partition:
  ID-1: / size: 232.43 GiB used: 31.47 GiB (13.5%) fs: ext4 dev: /dev/dm-0
    mapped: luks-a8eaaa90-b4ba-4943-8c1d-ddace5892f40
  ID-2: /boot size: 973.4 MiB used: 524.1 MiB (53.8%) fs: ext4
    dev: /dev/sda2
  ID-3: /boot/efi size: 252 MiB used: 274 KiB (0.1%) fs: vfat dev: /dev/sda1
Swap:
  ID-1: swap-1 type: file size: 3 GiB used: 3.8 MiB (0.1%) file: /swap/swap
Sensors:
  System Temperatures: cpu: 80.0 C pch: 68.0 C mobo: 48.0 C
  Fan Speeds (RPM): N/A
Info:
  Processes: 251 Uptime: 33m Memory: 7.65 GiB used: 3.56 GiB (46.6%)
  Init: SysVinit runlevel: 5 Compilers: gcc: 12.2.0 Packages: 2789 Shell: Bash
  v: 5.2.15 inxi: 3.3.26

17 *    * * *   root    cd / && run-parts --report /etc/cron.hourly
25 6    * * *   root    test -x /usr/sbin/anacron || { cd / && run-parts --report /etc/cron.daily; }
47 6    * * 7   root    test -x /usr/sbin/anacron || { cd / && run-parts --report /etc/cron.weekly; }
52 6    1 * *   root    test -x /usr/sbin/anacron || { cd / && run-parts --report /etc/cron.monthly; }

The 17" Laptop is running Linux Mint 22.1 (Cinnamon). This is the one that has the five minute chugs.

System:
  Kernel: 6.8.0-58-generic arch: x86_64 bits: 64 compiler: gcc v: 13.3.0
  Desktop: Cinnamon v: 6.4.8 Distro: Linux Mint 22.1 Xia
    base: Ubuntu 24.04 noble
Machine:
  Type: Laptop System: Dell product: Inspiron 7773 v: N/A
    serial: <superuser required>
  Mobo: Dell model: 0R58C3 v: A00 serial: <superuser required> UEFI: Dell
    v: 1.19.0 date: 12/15/2021
Battery:
  ID-1: BAT0 charge: 34.9 Wh (97.5%) condition: 35.8/56.0 Wh (63.9%)
    volts: 16.0 min: 15.2 model: Samsung SDI DELL W7NKD7B status: discharging
CPU:
  Info: quad core model: Intel Core i7-8550U bits: 64 type: MT MCP
    arch: Coffee Lake rev: A cache: L1: 256 KiB L2: 1024 KiB L3: 8 MiB
  Speed (MHz): avg: 658 high: 867 min/max: 400/4000 cores: 1: 400 2: 800
    3: 400 4: 400 5: 800 6: 800 7: 867 8: 800 bogomips: 31999
  Flags: avx avx2 ht lm nx pae sse sse2 sse3 sse4_1 sse4_2 ssse3 vmx
Graphics:
  Device-1: Intel UHD Graphics 620 vendor: Dell driver: i915 v: kernel
    arch: Gen-9.5 bus-ID: 00:02.0
  Device-2: NVIDIA GP108M [GeForce MX150] vendor: Dell driver: nvidia
    v: 550.120 arch: Maxwell bus-ID: 01:00.0
  Device-3: Realtek Integrated_Webcam_HD driver: uvcvideo type: USB
    bus-ID: 1-5:2
  Display: x11 server: X.Org v: 21.1.11 with: Xwayland v: 23.2.6 driver: X:
    loaded: modesetting,nvidia unloaded: fbdev,nouveau,vesa dri: iris gpu: i915
    resolution: 1920x1080~60Hz
  API: EGL v: 1.5 drivers: iris,nvidia,swrast platforms:
    active: gbm,x11,surfaceless,device inactive: wayland,device-2
  API: OpenGL v: 4.6.0 compat-v: 4.5 vendor: intel mesa
    v: 24.2.8-1ubuntu1~24.04.1 glx-v: 1.4 direct-render: yes renderer: Mesa
    Intel UHD Graphics 620 (KBL GT2)
Audio:
  Device-1: Intel Sunrise Point-LP HD Audio vendor: Dell driver: snd_hda_intel
    v: kernel bus-ID: 00:1f.3
  API: ALSA v: k6.8.0-58-generic status: kernel-api
  Server-1: PipeWire v: 1.0.5 status: active
Network:
  Device-1: Intel Wireless 3165 driver: iwlwifi v: kernel bus-ID: 02:00.0
  IF: wlp2s0 state: up mac: <filter>
Bluetooth:
  Device-1: Intel Bluetooth wireless interface driver: btusb v: 0.8 type: USB
    bus-ID: 1-7:3
  Report: hciconfig ID: hci0 rfk-id: 4 state: up address: <filter> bt-v: 4.2
    lmp-v: 8
Drives:
  Local Storage: total: 238.47 GiB used: 36.5 GiB (15.3%)
  ID-1: /dev/nvme0n1 vendor: Samsung model: MZVLB256HBHQ-000H1
    size: 238.47 GiB temp: 25.9 C
Partition:
  ID-1: / size: 229.63 GiB used: 36.21 GiB (15.8%) fs: ext4 dev: /dev/dm-1
    mapped: vgmint-root
  ID-2: /boot size: 1.61 GiB used: 291.7 MiB (17.7%) fs: ext4
    dev: /dev/nvme0n1p2
  ID-3: /boot/efi size: 511 MiB used: 6.1 MiB (1.2%) fs: vfat
    dev: /dev/nvme0n1p1
Swap:
  ID-1: swap-1 type: partition size: 1.91 GiB used: 0 KiB (0.0%)
    dev: /dev/dm-2 mapped: vgmint-swap_1
Sensors:
  System Temperatures: cpu: 30.0 C pch: 32.5 C mobo: N/A
  Fan Speeds (rpm): N/A
Info:
  Memory: total: 16 GiB available: 15.36 GiB used: 1.82 GiB (11.9%)
  Processes: 338 Uptime: 2h 38m Init: systemd target: graphical (5)
  Packages: 1996 Compilers: gcc: 13.3.0 Shell: Bash v: 5.2.21 inxi: 3.3.34

17 *	* * *	root	cd / && run-parts --report /etc/cron.hourly
25 6	* * *	root	test -x /usr/sbin/anacron || { cd / && run-parts --report /etc/cron.daily; }
47 6	* * 7	root	test -x /usr/sbin/anacron || { cd / && run-parts --report /etc/cron.weekly; }
52 6	1 * *	root	test -x /usr/sbin/anacron || { cd / && run-parts --report /etc/cron.monthly; }

Help Request

Anyone have any ideas for tracking down what might be causing this? I was going to just wipe the machine and replace Linux Mint with MX Linux to rule that out, but I figured I'd ask here before doing that, especially because it could be the hardware and not the distro that's causing the issue.

I think the answer is probably yes, but I wanted to double check. Does anyone here have experience with them? For clarity, this is the home modems that you can buy for stationary 5g internet. This is not those pocket, battery powered personal hotspots.

I just moved into a new studio in France. Despite literally every other building on the street having fiber, my studio does not have it. So I am stuck with DSL, since France doesn’t use cable for internet service. My dsl gets 2-10 Mbps down and 0.25-1 up. This is abysmal, especially the upload.

The other problem is the building is such a good shield against radio waves. With Orange, supposedly the best mobile carrier in France, I get exactly zero service, even sticking my phone out the window.

There is a carrier called Free, that does offer a 5g home internet box. To test it out, I got a Free (but not free) SIM card for my phone. Sticking it out the window, I get about the same speeds as my DSL box provides.

So there is the impetuous for my question. Should I expect their 5g home internet box to be about the same speeds as my iPhone, or can I expect them to be faster and more reliable, since it isn’t battery powered and can me larger?

Curious to know what museums people have visited that really left them thinking, "this was worth it" (time, money, whatever "worth it" means to you).

I will start.

Minneapolis, MN
Mill City Museum

This museum really shocked me with its breadth and depth! The location and setting are really cool as well. It has parts of the ruins of the historical building, integrated with some beautiful architecture of the new building.

The museum aims to explain how the grain industry was established and blew up in the Midwest, and what special role it had in shaping Minnesota. It goes all the way back to discussing WWII up to the advent of convenience foods like Betty Crocker and Pillsbury. There is even a Betty Crocker test kitchen in museum, complete with all the smells pumped into the air.

The displays are cool and I think are kid friendly, while also not boring the shit out of adults. My favorite part of the museum is the elevator, I truly believe it's one of the coolest displays(?) I've seen in a museum before, but I don't want to give away anymore, lest I make anyone's expectations too high and spoil it lol.

Curious to know if we have a current morphometric based system that can detect with the same accuracy as DNA testing, if two people are related, without a priori knowledge that they are related, if that makes sense.

Meaning, if a system is fed 100 random photos of humans, but is not told "there are definitely related people in here" can it match it as accurately as a DNA test of those same humans' DNA samples?

This topic is for the Three Cheers for Tildes mobile app.

I'll summarize the major updates at the start of each similar topic, so people can read the updates and then hit Ignore if they don't care about more frequent updates and user feedback.

Recently:

[Android] Version 1.4.3 (Apr 30, 2025): Fixed a layout bug on topics.

[Android] Version 1.4.2 (Apr 11, 2025): Reduced highlighting when formatting markdown. Fixed minor text size bugs.

[iOS] Version 1.4.1 (Apr 11, 2025): Fixed a bunch of text size bugs reported through TestFlight, especially when rendering comments. Reduced highlighting when formatting markdown.

Version 1.4.0 (Apr 6, 2025):

• Added text size setting
• Fixed markdown formatting bar bugs

The text size setting for accessibility is long overdue. I've been feeling bad that some users couldn't even use the app because the text was too small.

This has been another large change where I had to go back and re-test screens throughout the entire app, and fix many layout bugs caused by the dynamic text size. It's been very tedious!

In fact, the iOS release is delayed because I found some last-minute bugs and have had to go back to figure out solutions. iOS is up on TestFlight!

Also I am aware that there are still bugs in some places when you set the text excessively large. It's not a priority for me to fix those, unless they make the app unusable.

Have been particularly busy so far this year and that will continue for a while, so I may be less responsive here, even though I likely will see your messages. Thanks for continuing to report issues; v1.4 fixes some bugs based on those reports.

Previous topic: February 2025

Where to get it

Android version on Google Play Store: https://play.google.com/store/apps/details?id=com.talklittle.android.tildes

Or sideloadable APK at https://www.talklittle.com/three-cheers/

iOS version on the App Store: https://apps.apple.com/app/three-cheers-for-tildes/id6470950557

Join TestFlight for iOS beta testing: https://testflight.apple.com/join/mpVk1qIy

The popular online discourse on Large Language Models’ (LLMs’) capabilities is often polarized in a way I find annoying and tiresome.

On one end of the spectrum, there is nearly complete dismissal of LLMs: an LLM is just a slightly fancier version of the autocomplete on your phone’s keyboard, there’s nothing to see here, move on (dot org).

This dismissive perspective overlooks some genuinely interesting novel capabilities of LLMs. For example, I can come up with a new joke and ask ChatGPT to explain why it’s funny or come up with a new reasoning problem and ask ChatGPT to solve it. My phone’s keyboard can’t do that.

On the other end of the spectrum, there are eschatological predictions: human-level or superhuman artificial general intelligence (AGI) will likely be developed within 10 years or even within 5 years, and skepticism toward such predictions is “AI denialism”, analogous to climate change denial. Just listen to the experts!

There are inconvenient facts for this narrative, such as that the majority of AI experts give much more conservative timelines for AGI when asked in surveys and disagree with the idea that scaling up LLMs could lead to AGI.

The ARC Prize is an attempt by prominent AI researcher François Chollet (with help from Mike Knoop, who apparently does AI stuff at Zapier) to introduce some scientific rigour into the conversation. There is a monetary prize for open source AI systems that can perform well on a benchmark called ARC-AGI-2, which recently superseded the ARC-AGI benchmark. (“ARC” stands for “Abstract and Reasoning Corpus”.)

ARC-AGI-2 is not a test of whether an AI is an AGI or not. It’s intended to test whether AI systems are making incremental progress toward AGI. The tasks the AI is asked to complete are colour-coded visual puzzles like you might find in a tricky puzzle game. (Example.) The intention is to design tasks that are easy for humans to solve and hard for AI to solve.

The current frontier AI models score less than 5% on ARC-AGI-2. Humans score 60% on average and 100% of tasks have been solved by at least two humans in two attempts or less.

For me, this helps the conversation about AI capabilities because it gives a rigorous test and quantitative measure to my casual, subjective observations that LLMs routinely fail at tasks that are easy for humans.

François Chollet was impressed when OpenAI’s o3 model scored 75.7% on ARC-AGI (the older version of the benchmark). He emphasizes the concept of “fluid intelligence”, which he seems to define as the ability to adapt to new situations and solve novel problems. Chollet thinks that o3 is the first AI system to demonstrate fluid intelligence, although it’s still a low level of fluid intelligence. (o3 also required thousands of dollars’ worth of computation to achieve this result.)

This is the sort of distinction that can’t be teased out by the polarized popular discourse. It’s the sort of nuanced analysis I’ve been seeking out, but which has been drowned out by extreme positions on LLMs that ignore inconvenient facts.

I would like to see more benchmarks that try to do what AGI-AGI-2 does: find problems that humans can easily solve and frontier AI models can’t solve. These sort of benchmarks can help us measure AGI progress much more usefully than the typical benchmarks, which play to LLMs’ strengths (e.g. massive-scale memorization) and don’t challenge them on their weaknesses (e.g. reasoning).

I long to see AGI within my lifetime. But the super short timeframes given by some people in the AI industry feel to me like they border on mania or psychosis. The discussion is unrigorous, with people pulling numbers out of thin air based on gut feeling.

It’s clear that there are many things humans are good at doing that AI can’t do at all (where the humans vs. AI success rate is ~100% vs. ~0%). It serves no constructive purpose to ignore this truth and it may serve AI research to develop rigorous benchmarks around it.

Such benchmarks will at least improve the quality of discussion around AI capabilities, insofar as people pay attention to them.

Update (2024-04-11 at 19:16 UTC): François Chollet has a new 20-minute talk on YouTube that I recommend. I've watched a few videos of Chollet talking about ARC-AGI or ARC-AGI-2, and this one is beautifully succinct: https://www.youtube.com/watch?v=TWHezX43I-4

As per title this is inspired by my recent problems with a Latitude 7320 notebook.

I can't use my desktop right now and so wanted some cheaper nb for normal usage and eventually settled on this model due to being able to get it at an acceptable ratio of price to age and seeing it as compatible on Ubuntu, not noticing the disclaimer until later.

The problems started right after installing Fedora KDE - the nb was running at absolutely abysmal performance and this problem affects several models.

Running passmark I've got above 2000 on cpu, on Windows I had 11000. The cpu was throttling to 1500Mhz and lower for no reason. Switching a BIOS setting of power management to "ultra performance" got me to twice the score.

Eventually using throttled from github for various Lenovo and Dell models and thermald I was able to get to twice that again, still a fifth less than on Windows. Also the repo has potential of security concerns due to how it works, also potential to just stop working due to them later.

Mainly I'm posting this to just say that there is zero legitimate technical reason why this should happen, it works on Windows and on Dell tampered Ubuntu images. The hw is fine but for some reason someone somewhere decided to artificially limit the hw for whatever reason.

Right now I am still indecided if I should write off the several hours I've spent on this and return the machine to play the dice with some other model.

Edit 5.4.: it turns out I was not using the throttled package correctly and now have roughly equivalent performace in Linux as in Windows up from the 4/5 or so after all the other workarounds. All of the points still apply though. I also heartily recommend s-tui as a nice utility for cpu monitoring and stress test.

I'm wondering if a post that has tags that I've filtered will still generate notifications, so I've filtered the tag kfwyre.testing and then applied it to this post.

If someone seeing on this could comment on it so that I can see if I get a notification for it, please do so!

The AMD Radeon RX 9070 XT launches tomorrow (6th March) for a MRRP of:

• $600 USA before taxes
• £570 UK after taxes

Reviews:

• Have They Finally Done It? - Hardware Unboxed
• Review & Benchmarks vs. 5070 Ti, 5070, 7900 XT - Gamer Nexus
• Nvidia in Trouble? The RX 9070 XT has great potential - der8auer
• AMD Radeon RX 9070 / 9070 XT review: back to winning ways - Eurogamer
• AMD, I Could Kiss You - 9070 and 9070 XT Review - Linus Tech Tips
• AMD's Radeon RX 9070 XT is the most exciting GPU to launch in years - XDA Developers

Bonus:

• AMD FSR 4 Upscaling Tested vs DLSS 3/4 - A Big Leap Forward - RDNA 4 Delivers! - Digital Foundry