Why Do I Keep Getting Called Out - WAN Show December 2, 2022

Visit www.squarespace.com/WAN and use offer code WAN for 10% off
Check out Newegg at newegg.io/linustt2022
Start building your credit with debit today at www.extra.app/wanshow
Extra is a technology company. The Extra Debit Card is issued by Evolve Bank & Trust, Member FDIC, pursuant to a license from Mastercard USA. Extra reports on time and late payments, which may negatively impact your credit score. Credit scores are independently determined by credit bureaus based on a number of factors, including your other financial transactions. Extra reports to Experian® and Equifax®. Please refer to extra.app/policies for additional details.
Podcast Download: spotifyanchor-web.app.link/e/...
Timestamps: (Courtesy of NoKi1119)
0:00 Chapters
2:09 Intro
2:38 Topic #1 - Updates on the eufy situation
3:46 Rob's video, notifications, port-forwarding & 30-day deletion
5:14 Data notification, Linus on heat pumps & Luke corrections
8:24 Users information, GDPR, event logs & EULA
12:24 taping over lights
14:09 How would you provide "free" local storage?
14:56 eufy's quote, Luke's hot take
16:12 Linus on hand-picking data & "local-only"
20:06 Topic #2 - OpenAI's ChatGPT text & code generator
25:02 ChatGPT, security filters, Luke's ballad on butts
31:32 Linus unable to log into beta, discusses prompt
34:36 Linus's prompt on Luke's butt
35:54 Luke uses his character & prompts stories via chatbot
37:04 Discussing Advent of Code, automated prompts
40:42 OpenAI prompted a warning to Linus
42:28 Learning programming via chatbot, API docs reference
44:54 Broken GPU scenario, generating an LTT video script
48:38 Chatbot's advice on feeding kids vegetables
50:56 Chatbot's spicy dbrand Tweet, Luke on jobs future
56:12 Chatbot explaining world problems Kanye West's style
57:22 Sponsors ft. OpenAI
57:54 Sponsor - Squarespace
59:26 Sponsor - Newegg
1:01:54 Sponsor - Extra Credit
1:04:06 Linus tries the chatbot for Extra again
1:05:28 Compiler error in generated code, API issues
1:07:24 LTTStore limited amount of black shaft screwdrivers
1:09:22 Topic #3 - LTT Labs beta site demo
1:12:52 Preview of comparing two product reviews
1:14:14 LMG still hiring
1:14:55 Topic #4 - Apple's SOS saves a stranded Snowmobiler
1:16:28 Apple shuts off AirDrop in China after 10 minutes
1:18:08 ChatGPT makes Discord bot for responses, poor Jake
1:18:48 Topic #5 - The Last Seven Days in Twitter #3
1:19:26 Elon on Apple not advertising & "free speech suppression"
1:20:02 Hunter Biden story suppression
1:20:18 Elon on an "alternate phone," Linus calls everyone out
1:20:48 No longer enforcing COVID misinformation policy
1:21:08 Elon resolves misunderstanding with Tim Cook
1:23:22 Why not make a fork? Discussing Logitech G Cloud
1:25:58 Twitter engagement, trending & business aspect
1:31:10 Linus calls out FP user on "children consent"
1:31:56 Topic #6 - DnD will now call races "species"
1:33:18 Merch Messages #1
1:33:29 Encouragement for those entering the developer space
1:49:38 Do over-regulations stifle innovation?
1:51:18 With similar AI threads, would it be unique?
1:54:12 Topic #7 - Scalpers unable to sell RTX 4080 or return it
1:54:40 Hardware Unboxed's video on a "new" RTX 3060 8GB
1:56:14 Comparing relative performance, low-end alternatives
1:59:38 Merch Messages #2
1:59:46 How would Luke spend LMG's money?
2:02:46 Topic #8 - Luke on Intel's Arc
2:07:33 Would people overlook AMD's flaws after Arc's introduction?
2:09:38 Luke discusses AV1 for streamers, AV1 advantages
2:11:28 Would Linus put an RDNA3 GPU in his personal rig?
2:12:24 Luke is not sure what to do with his money
2:13:17 Topic #9 - Hacksmith's "fire-cooled" PC collab
2:13:55 Showcasing Mini-Saber
2:16:17 Reselling PTM7950 idea, conflict of interest & assets
2:20:22 Validating data, transparency, trust & spoofing
2:23:38 Merch Messages #3
2:23:46 Thoughts on the removal of product pages
2:24:56 Using DXVK, Valve's Proton
2:27:01 What's next on a future home tour?
2:27:46 Selling branded merch to fund open source projects?
2:30:03 Outro

[@NoKi1119]

Bring it on, ChatGPT - there can be only one timestamper! -Timestamps- [0:00] *Chapters.* [2:09] *Intro.* [2:38] *Topic #1: Updates on the eufy situation.* > 3:46 Rob's video, notifications, port-forwarding & 30-day deletion. > 5:14 Data notification, Linus on heat pumps & Luke corrections. > 8:24 Users information, GDPR, event logs & EULA. > 12:24 taping over lights. > 14:09 How would you provide "free" local storage? > 14:56 eufy's quote, Luke's hot take. > 16:12 Linus on hand-picking data & "local-only." [20:06] *Topic #2: OpenAI's ChatGPT text & code generator.* > 25:02 ChatGPT, security filters, Luke's ballad on butts. > 31:32 Linus unable to log into beta, discusses prompt. > 34:36 Linus's prompt on Luke's butt. > 35:54 Luke uses his character & prompts stories via chatbot. > 37:04 Discussing Advent of Code, automated prompts. > 40:42 OpenAI prompted a warning to Linus. > 42:28 Learning programming via chatbot, API docs reference. > 44:54 Broken GPU scenario, generating an LTT video script. > 48:38 Chatbot's advice on feeding kids vegetables. > 50:56 Chatbot's spicy dbrand Tweet, Luke on jobs future. > 56:12 Chatbot explaining world problems Kanye West's style. [57:22] *Sponsors ft. OpenAI.* > 57:54 Squarespace: all-in-one website platform. > 59:26 Newegg: leading online retailer. > 1:00:55 Newegg: Benefits for holiday shopping. > 1:01:54 Extra "Gum" Credit. [Cont.] *Topic #2: OpenAI's ChatGPT.* > 1:04:06 Linus tries the chatbot for Extra again. > 1:05:28 Compiler error in generated code, API issues. [1:07:24] *LTTStore limited amount of black shaft screwdrivers.* [1:09:22] *Topic #3: LTT Labs beta site demo.* > 1:12:52 Preview of comparing two product reviews. [1:14:14] *LMG still hiring.* [1:14:55] *Topic #4: Apple's SOS saves a stranded Snowmobiler.* > 1:16:28 Apple shuts off AirDrop in China after 10 minutes. > 1:18:08 ChatGPT makes Discord bot for responses, poor Jake. [1:18:48] *Topic #5: The Last Seven Days in Twitter #3.* > 1:19:26 Elon on Apple not advertising & "free speech suppression." > 1:20:02 Hunter Biden story suppression. > 1:20:18 Elon on an "alternate phone," Linus calls everyone out. > 1:20:48 No longer enforcing COVID misinformation policy. > 1:21:08 Elon resolves misunderstanding with Tim Cook. > 1:23:22 Why not make a fork? Discussing Logitech G Cloud. > 1:25:58 Twitter engagement, trending & business aspect. > 1:31:10 Linus calls out FP user on "children consent." [1:31:56] *Topic #6: DnD will now call races "species."* [1:33:18] *Merch Messages #1.* > 1:33:29 Encouragement for those entering the developer space. > 1:49:38 Do over-regulations stifle innovation? > 1:51:18 With similar AI threads, would it be unique? [1:54:12] *Topic #7: Scalpers unable to sell RTX 4080 or return it.* > 1:54:40 Hardware Unboxed's video on a "new" RTX 3060 8GB. > 1:56:14 Comparing relative performance, low-end alternatives. [1:59:38] *Merch Messages #2.* > 1:59:46 How would Luke spend LMG's money? [2:02:46] *Topic #8: Luke on Intel's Arc.* > 2:07:38 Would people overlook AMD's flaws after Arc's introduction? > 2:08:51 Luke discusses AV1 for streamers, AV1 advantages. > 2:11:28 Would Linus put an RDNA3 GPU in his personal rig? > 2:12:24 Luke is not sure what to do with his money. [2:13:17] *Topic #9: Hacksmith's "fire-cooled" PC collab.* > 2:13:55 Showcasing Mini-Saber. > 2:15:52 Reselling PTM7950 idea, conflict of interest & assets. > 2:20:22 Validating data, transparency, trust & spoofing. [2:23:38] *Merch Messages #3.* > 2:23:46 Thoughts on the removal of product pages. > 2:24:56 Using DXVK, Valve's Proton. > 2:27:01 What's next on a future home tour? > 2:27:46 Selling branded merch to fund open source projects? [2:30:03] *Outro.*

[@Allibabo]

Hooow?!

[@crowlsyong]

Wow man. Legendary. So fast.

[@rustyshackleford5166]

@@Allibabo maybe he used the AI? lol

[@NotJohnnyTamale]

Lmao

[@mahjonglegendISLAM]

That's frickin awesome

[@carston101]

I would absolutely watch a full 2hr+ podcast dedicated to you guys messing with the chat bot. I hadn't even heard about prior to this but what you guys demonstrated here was just insane.

[@sligit]

I can highly recommend the several Corridor Crew podcasts where they just play with GPT3.

[@carston101]

@@sligit on KZbin or their site?

[@sligit]

@@carston101 KZbin. I think it's called corridor cast, on the corridor crew channel iirc.

[@carston101]

@@sligit corridor cast is it. Damn, been watching their other channel since the very first VFX artist react, but had no idea they did a podcast. Thanks!

[@sligit]

@@carston101 Yeah I only discovered it fairly recently too. There's lots of good deeper discussion on the about topics they cover in their other videos.

[@NoHope_]

I love when luke is super passionate about something he's talking about

[@brightshadowdenmark]

i would definitely watch a 4 hour dedicated livestream of Luke, Linus, Alex, Jake and James playing with Open AI :D Edit: woops forgot Anthony on my AI- Dream Team ;-)

[@kisspeteristvan]

i would never watch that , i'm not 15 anymore and wasting 4 hours straight just doesn't sound right .

[@brightshadowdenmark]

@@kisspeteristvan LOL.. (279 Likes, and only one negative comment)

[@JAVIERsoul]

@@kisspeteristvan I mean you wouldn't have to watch it all in one go

[@zenot1c237]

@@kisspeteristvan id just put it on in the background

[@thumbtak123]

@@kisspeteristvan unless you are like me with too much free time and no one to spend it with.

[@haydenb123]

Users: "Eufy, you lied in your marketing!" Eufy: "I reject your reality and substitute my own!"

[@tonn333]

Replace Eufy on the last line with "most of the world" Comforting lies are much more popular than the reality. In this crazy "post truth" age (as they say), most think they can have "their truth".

[@rtyler1869]

Sounds a lot like a certain wing of a certain USA political party

[@mrblackrock555]

Please don't taint the words of the savage that is the Adam by bring him into the eufy situation please :(

[@joshjlmgproductions3313]

"Reality can be whatever I want it to be."

[@haydenb123]

Haha love these replies

[@davidnotonstinnett]

You cannot contract illegal activity, at least not in the US. Just because you sign a contract does not give someone the ability to break the law against you.

[@CanIHasThisName]

Exactly. I also believe that EULAs can (in some cases) become unenforceable if you can only agree with them after purchase or if their terms are unreasonable. And I think there's definitely an argument to be made that if a product tells you in its marketing and on the package that it doesn't do something, the EULA making you agree it can do that would be unreasonable. And as you pointed out, if it's trying to make you agree to something it cannot legally do, it's not valid.

[@Orochistorm]

@@CanIHasThisName The vast, vast majority of EULAs are actually wholly unenforcable and do not hold up in court. There is an immense precedent for both judges and juries ruling that "no one fucking reads that shit, stop sneaking evil garbage into it" is the defacto response to any suit using a EULA as its basis.

[@Aidiakapi]

​@@CanIHasThisName And EULAs cannot override laws anyways. This whole ordeal is surely breaking both marketing and privacy laws, at least here in the EU.

[@Pleezath]

Also when it keeps updating like Samsung's terms of service for their phones. Like you already have it for at least 1 to 2 years and the EULA / terms of service keeps changing constantly. Like I can't stop using my phone if I disagree because it is so integrated into our life's

[@XionEternum]

Tell that to the US Military. I found out that someone who enlisted got sunburnt and was then sued for damaging government property... because they were government property and willfully damaged their body in an act of negligence.

[@shApYT]

We need a 2 hour long ChatGPT live stream with Linus and Luke.

[@adarshjkalathil]

Everybody like this so Linus see's this

[@ChristianStout]

You can do that by yourself at home: open two ChatGPT threads in different browsers, have one play the role of Luke, and the other play the role of Linus. Copy/paste the reply from one into the prompt of the other, repeat ad nauseam.

[@SuspiciousFish144]

@@ChristianStout this is what people mean when they say AI won't be replacing people🤣

[@onlypoynter]

This chatGPT is absolutely insane. This thing is just going to absolutely revolutionize so much. To be completely honest this is genuinely scary

[@shadowboyii]

And as far as we know it is just the beginning

[@Marquuus]

The number of things I've thrown at it and it has excelled at is beyond awe-inspiring.

[@pleepler]

@@Marquuus I threw some stuff at it yesterday and the more complex the prompt you give it, the worse the result gets. Usually it agrees with you, which is kinda dumb, it just repeats what you say and adds something to it. But if you give it a short, pointy prompt, it can give you nice results I've experimented a little bit more, and it takes a lot of going back and forth to get the result you actually want. I make the AI do maths and the number of times it forgot some parameters I set is annoying. Eventually we got to a result I found correct, but it did take a lot of effort and I had to explain like it's a child I'm talking to

[@StormShifter]

I’m just glad I don’t have to do my essay now lol I got it to do it and it got an A 😂

[@joostverra9130]

Yeah it's gonna be fun in the beginning, until it's not anymore...

[@Duny645]

I could listen to Luke talk about the chatbot for hours

[@OperationDarkside]

Maybe we should bully him into doing just that.

[@phenomanII]

I am used to watching the WAN show for an entertaining look at the news with some tangents about something unrelated - the more of those, the better. I've had my mind blown by some things in the past, but mostly by how poorly something was implemented or how poorly a company responded to an issue, much like Eufy in this one. The ChatGPT segment on the other hand completely flabbergasted me. I don't even know if I can put all these feelings into words, but I bet *it* could. Insane.

[@SegmentW]

I hope they talk more about it next WAN show. I really do

[@AnthonyChopra]

Luke : it can fix stuff !!

[@littlegreenman91]

Absolutely. I got hooked on the chatbot after this segment. And well.. I was able to let it write me some code within 30 minutes that took me half a day. This is gonna be so damn useful. Not even in the future. But right now!

[@stanhilt1912]

I just asked the chatbot to write me a short assay about an incredibly specific antigen for a neglected tropical disease. Not only did it get 95% of the information right, it also managed to include it's possible future uses. I am beyond impressed and slightly concerned

[@elknackebroto7447]

There is no need

[@lasthopelost9090]

That what I’m wondering what can this come up with like can it come up with things like cures we know it can make code and fix the error when asked too crazy

[@rpavlik1]

All it's doing is predicting the next word based on the words it already has and the training data.

[@3rgoproxxy]

You just asked it about something that's already in it's database. It's the equivalent of repacking information from google in a more easily accessible way.

[@OfficialSamuelC]

@@3rgoproxxy That’s the point though. He knows it’s available to the public. But it could take hours crawling search engines and dozens of different sites to find everything you need on something. Having this bot just have everything in one place is amazing and convenient. Nobody is saying it’s coming up with new things. It’s just amazing how convenient and accurate it is. Amount of time it would save. Even if you added the time to verify things, it’s still quicker than manually doing it all.

[@sipos0]

Using it for the sponsor talking points is genius. I have never listened to a SquareSpace advert so closely.

[@youtubeaddict9393]

I didn’t realize how crazy the AI was until Linus used it to write a script instantly. That’s nuts.

[@TheSamDickey]

I’m watching this at 3:30 am because my neighbors are f****** very loudly and I want to drown out the noise. WAN show is a life saver

[@SwervingLemon]

I forgot apartments exist while I was reading this. Thought briefly that your neighbors were very impressive.

[@TheSamDickey]

@@SwervingLemon Yep just moved out into my first apartment in July. Cannot wait to purchase a house 😃. Hopefully the recession hits hard and house prices dip in about a year lol

[@Fabio53443]

Wan(k) show Sorry

[@PrestoJacobson]

This is so funny imagining you listening to this show, with them still loud in the background. I'd still hear them for sure unless I used earplugs/+music.

[@Mradevans]

@@TheSamDickey depending on locale, theyve dipped about 25% asking price from march to nov

[@harbirsingh7266]

OpenAI GPT3 is SCARY. I knew this was coming, but didn't think it would happen THIS FAST. Writing correct code instantly is MUCH MUCH MORE impressive than most people realize.

[@WheatGrinding]

Is this what John Henry felt as he was watching the steam engine get delivered?

[@Ardeact]

honestly it's best to ignore it and wait until our inevitable demise with it, there's nothing we really can do

[@GigaBrand]

Right?!? And if you tell it to write its code for itself… we’ll that’s a self improving algorithm…basically one step from a General AI intelligence

[@bluelightfox]

Especially when that code the AI has 'written' closely matches what someone else wrote, say under an open source license, but without informing you of the license. There's a large class action lawsuit against GitHub over its CoPilot, which was 'trained' on open source projects. The litigation that proceeds from this will likely also be 'scary.'

[@jasondrummond9451]

To survive we may have to go 'low tech' Battlestar Galactica style - NO computer control of ANYTHING. This is the Mother of all Paradigm Shifts.

[@shishsquared]

When I first heard of this tool, I was like "eh whatever. " And WAN started with Luke going "I was working with a chatbot for several hours" I was like "A little weird, but okay" then after actually seeing this . . . yeah I'm convinced this is the next big leap. Luke is exactly right about how this is just as impactful if not moreso than Google was when it first released.

[@christophergilbert5988]

Honestly, the effects of GPT3 on the education system might be very good. Because there is now an AI that can do most text-based homework for you, schools may change their approach to make learning more experiences-based and interactive, which could turn out to be a very good thing.

[@theexaustedslime]

I hope you're right. I suspect otherwise.

[@GianniLeonhart]

lmao he thinks this wont be abused and that the education system will keep up

[@ThePaintitwhite]

As part of the education system (though not in the US) I can totally see some teachers trying to change this and then having admins complain and parents complain and then giving up on it... Changing the education system is hard, they love their old ways and the blame falls on you as a teacher for little reward so it's hard to stay motivated. I don't want to be so negative though. I've been thinking of awesome ways to use it so there is indeed a lot of potential and I'd love to include it somehow. I've already tried using it to make some things faster so I can focus on lesson planning though it's not very good at giving feedback, it just says everything is wrong and gives really strange suggestions, just to then use the original thing as a "correct" example.

[@colelangford6369]

Optimistic, but that would be nice.

[@justsomeguywithnoeyes7303]

i guess future generations of kids will be dumber and dumber and more reliant on AI / technology to do the thinking for them

[@JulienChabanon]

Luke's reaction when Linus read the SquareSpace sponsor spot generated by the AI is priceless!!!!

[@samch.8351]

I dont understand how after spending many hours with it hes so impress at the AI basically just repeating what we hear Linus say with every squarespace sponsor when the thing wrote him some actual code lol

[@RedOnRye]

@@samch.8351 he didn't try to get GPT emulating a specific person in script format, so he was impressed with how well it did.

[@samch.8351]

@@RedOnRye Yes exactly, but again I dont understand that after spending so many hours with something that gave him way more "advance answer" many times hes impress by the thing basically doing a copy/paste

[@Naruto31132]

@@samch.8351 so you view it like a buggy search engine or futuristic slightly inaccurate encyclopedia. I get it, it's similar to Google with the answers, just decluttered and no ads (for now) with a bit of RNG and chance of inaccuracy to it. I'm still very impressed with it, as a type of knowledgebase medium, at worst, given the margin of error.

[@KingLarbear]

IA or AI

[@thelongwinter362]

I just spent 3 hours messing with open AI gpt3 and I should have been studying for my chemistry final. Thanks linus

[@quackaduck42]

Why not use it for studying? You can compare notes. Might spew more advanced stuff than what you are studying but it can help. Also a chance to spew incorrect stuff because AI but it would help.

[@RadioactiveBlueberry]

@@quackaduck42 "Explain theory X in a way that a 10 year old can understand it" might be worth a try. Easy to read, easy to verify.

[@AnthonyChopra]

Luke : it can fix stuff !!

[@thelongwinter362]

@@RadioactiveBlueberry I have not tried a prompt like that yet. Luckily I think I can understand this stuff on my own. I am in my first year of engineering school at texas A&M and they make you take chemistry for some stupid reason. I don't even need chemistry in my desired major (computer engineering). Chem makes my blood boil but I have a still have a shot at an A

[@NoNameAtAll2]

make ai write your chemistry final for you

[@Cossieuk]

Dont just list the results from the lab, list the manufactures claims as well so people can see who accurate they are

[@666Tomato666]

For the lab, yes, probably publishing exact test procedures, including the tooling used, is really the best thing you can do.

[@YTMichaelFromTexas]

Eufy lost all "benefit of the doubt" by advertising "what happens local, stays local" and using that as a SELLING POINT!! I had been tempted to buy their products previously. Now? NO! FREAKIN! WAY! Dead to me as a company, now!

[@saladgreens912]

Genuinely unacceptable. I wish I could find a more strongly worded way to state that but it's just unacceptable. What sucks is that Eufy is a brand owned by Anker. I just bought an Anker USB hub a week before this came out. Had the timing been different I would have not. If Anker is okay with this behavior, I'm not okay with buying Anker products.

[@toine512fr]

Yeah, however cloud use is justified or not, they LIED. Criticising Linus position with technical arguments is completely out of touch.

[@shortboard_89]

I have one of their baby monitors which thankfully isn’t an internet connected device. No way I’d buy anything else now.

[@zqzj]

They're CCP spy devices. It wasn't a mistake. This was intentional.

[@ModrunOfficial]

what to buy instead? all the other ones seem to require subscription or wifi always on

[@tyw877]

Ok so I immediately created an OpenAI account, and dbrand works first try Prompt: write a satirical tweet from dbrand Result: "Introducing the new dbrand skin for your phone: the 'cracked screen' edition. Because who needs a case when you can just pretend your phone is already broken?"

[@GigaBrand]

Omg that is actually funny af what a time to be alive.

[@EverythingWentSideways]

been messing around with GPT3 all day because of this WAN show. it's got some interesting applications as a tool for writers block, busted out a few half finished song lyrics that i couldn't figure out where to go with and after messing around with them it really gave me some decent ideas. definitely not a fool proof "write me a song" engine or anything, but it certainly gave me some interesting phrases and concepts that i don't think i would have considered without it. very impressed.

[@PatrikHlebecStor]

Every week it's a delight to watch the show. It truly is a highlight of the week and I'm always excited about it.

[@AustinNooe]

Someone did a £300 super chat............. don't they say all the time they ignore those?

[@CakePrincessCelestia]

Should have gotten Backpack instead with a Merch Message! This way, 90 Quid went straigt to the gutter called Google! XD

[@wenziz87]

I think it’s less ‘ignore’ and more ‘don’t see’. I think it was said that their interface doesn’t highlight or filter superchats so they can’t curate them for responses.

[@sindicta5757]

A fool and his money are easily parted

[@kenosabi]

@@sindicta5757 he was lucky to ever have it to begin with

[@mon94key]

They say to buy merch because Google will take a 30% cut and by buying merch LTT gets more money and you get some swag in return along with a chance for the merch message to be seen so it's a win win for everyone

[@jeremymoller1038]

This AI is actually insane. I took it through a simple mechanical design brainstorm about a vacuum chamber and its responses were astonishing, you can get yourself a lot of the way there without having to sit there googling a bunch of things like I normally would for these kinds of sessions. Obviously though you will have to check your work, don't trust an AI to do the math on these things, double check. But wow the potential for this is INSANEEEE.

[@atienzo98]

Wild! The J.A.R.V.I.S to Tony Stark.

[@kennytvn]

The whole openai chat topic was super interesting, the extra time spent discussing it was well spent

[@jamalbarnes4327]

I love the "Labs tested" Idea for the store. Things we've seen you guys used on the channel, from other companies in your store.

[@seifenspender]

I feel like once I finally graduate as a computer scientist, AI/ML will have taken over nearly all of my interest fields.

[@ac.creations]

better start learning how to code with gpt3

[@JoppedeZeeuw]

I’m graduating as a mechatronics engineer, and let me tell you: as long as you can program even the slightest bit (and I’m talking like arduino-level programming), you will not be out of work. Most code I’ve come across when working on embedded systems is on the verge of breaking apart. When something works most of the time, it’s ready for market as long as you’re not working at a multi-billion dollar international firm like Philips, Siemens and the likes. Sp don’t worry, you won’t be out of work 😊

[@seifenspender]

@@ac.creations I threw my last years exam at it and it nearly got a perfect score. Wild stuff.

[@ruukinen]

@@JoppedeZeeuw Problem is that the ML stuff is getting into the programming field as well. Obviously I don't think ML will rob us of our jobs. I just think it will change how we work.

[@costafilh0]

Use it in your favor to go beyond! No point to ignore or trying to beat it.

[@expertb6348]

I wish Luke would do a stream about the AI chat bot. It was very interesting and in a weird way scary. I have 0 programing skills, but I would love to know how to use it.

[@GigaBrand]

YESSS!

[@kg4wwn]

Historically "We can do more work now, it won't cost jobs" has ALWAYS led to more money to the owner, less money to the workers.

[@djstringsmusic2994]

Yup always benefits the shareholders never the workers.

[@NoProHarrie]

This is an absolutely amazing episode! The rabbithole you guys went in to on this is absolutely amazing!

[@Apersonl0l]

The openAI story is some serious Human need not apply shit This is both terrifying and cool as hell

[@sid_the_ant]

Until AI take over my job, my job as a programmer has officially become much easier. Writing small scripts is no longer needed. Impressed and scared at the same time

[@jorismak]

Your job as a programmer is to make ai do your job, and double check it to see if its actually correct .

[@atienzo98]

@@jorismak amazing!!! I asked it to write a basic calculator, then I asked it to make it again but with a bug built in. It did it and pointed it out for me before I thought to ask.

[@ElGnacko]

I'm a technical translater and I remember back when Google translate was introduced how everybody said I'll be without a job soon. what actually happened was that machine translation increased my output tremendously and all the amateur wannabe translators were replaced by machines. Nowadays I spent most time correcting and proofreading machine translations. And this will happen for programmers too. The smart and professional people will keep their jobs, but all this van-life semiprofessional self-taught people will be gone.

[@ruukinen]

@@ElGnacko I mean you have to actually get the experience to become a professional at some point. There will just be van-life semi-professional ML prof programmers.

[@HansDenuevo]

@@ElGnacko but this any person in progress of being a professional coder/translator is losing their chances on making an income with their experience until they actually become a professional. The tools are convenient but you gotta admit it does cause problems to not fully realized workers. And that's how it's gotta go, the more jobs are replaced by AI, the bigger the gap will be for people to jump. Either you spend a lot of money you don't have to get a specialization or you earn no money because every other job is made by an AI/robot supervised by a professional now.

[@marcelroodt]

I share Luke's enthusiasm for the OpenAI chatbot. I think we're in a very interesting world right now. We knew that ML and AI have been getting more advanced, and hardware for them has similarly been improving. We now live in a world where you could use a chatbot to write scripts to impersonate people, Deepfake their voices and also generate pictures from prompts. The next step may just well be AI-generated video content. Great!

[@invertexyz]

Yes, and each step removes the need for humans more and more. At the end of the day what will we be left with? Just monkeys to consume, not really being important for creation anymore, no need for human-human collaboration. It's not a great future. This is why we need to try implementing laws around how these things source their data so that nothing short of a true artificial human intelligence can actually replace a human. AI should be used to do the kind of work WE DON'T WANT TO DO, not the work we are passionate about. Nobody wants to just become a "cleanup slave".

[@justsomeguywithnoeyes7303]

why would anyone be enthusiastic about AI impersonating people?

[@JeremieBPCreation]

I have butterflies about ChatGPT. Hadn't seen Linus looking that stunned in a while.

[@patrickbourne3819]

Okay that open AI is crazy! I just had a legit therapy session with it. Not saying that it's a replacement for an actual therapist but it's able to let me look up different coping mechanisms, ask dumb questions and go at my own pace without a timer.

[@justsomeperson5110]

I want Luke's Intel Extreme Tech Upgrade to be Luke using the money to hire/collaborate with Hacksmith to make a working hammer+crowbar design called the Lafbar, which LTT can find a manufacturer for to sell LTT+Hacksmith-branded Lafbars on the LTT store. Oh, and us software engineers have been jokingly predicting that we'd write AI to take our jobs from us for decades now. FINALLY the joke is becoming reality. I'm kind of scared to lose my job, but kind of excited to see how good it can run and how quickly it improves in years to come.

[@Haplo699g]

You mean a Halligan Bar?

[@coffilover]

He can't have one, because he's not a LTT employee but a Floatplane one. Stated it in a before video.

[@TroublesomeOwl]

@@coffilover And they stated on this wan show that it has been expanded to all branches

[@JaguarInfinity]

I've been trying to make my smart home as local as possible, but I have a Eufy doorbell. I agree the claims are somewhat misleading but I picked it for specific reasons. The data is stored locally without subscriptions which is what I wanted. I was aware there had to be data going over the cloud because the app connects to my device over the internet and I receive notifications over the internet. I was aware they were trying to improve on this but they hadn't managed to do far. Again I agree that the marketing is somewhat misleading, so I get why people are pissed, but I still prefer them to 100% cloud based services that continually charge me even if they are only 95% local

[@asineth832]

I would absolutely buy parts on LTT store if I could see real pictures of the product and maybe a Cinebench or 3DMark score on the page. That would be awesome!

[@ScottWinterringer]

My data is mine. And local means local. Why do people defend this... It's even worse when people double down and I literally dug into this.

[@ScottWinterringer]

@@adamk.7177 I Know how it works. This argument still makes no sense. Especially as it can't define whether the data is local or not. I can literally stream my own blueiris system or whatever rtsp from my own items...

[@weirdguybr]

@@adamk.7177 You mean the video where he states he is not a GDPR expert, but still proceeds to make claims about it, followed by incorrect technical claims about how notifications, encryption and CDNs work and ignores every single misleading marketing claim under an "implied consent"/"users should know better" approach? As a professional IT person who works exactly in the fields he made misleading claims about, I have to side with Linus and Luke here. Eufy screwed up and they are doing a horrible job at handling this; I'm not sure why Rob decided to jump on their side, but he is extremely wrong.

[@ScottWinterringer]

@@adamk.7177 yet its not.... and as we know there is a public portal with multiple users.....

[@ScottWinterringer]

@@adamk.7177 I am again also going to ask why this is being states as local and it is not

[@Kosmicseven]

Linus fart @ 2:03:00 Luke's reaction 😂

[@RafaelSoares89]

I was on headphones and that was hilarious, had to go back to make sure tho. I'd have aknowledged it if I were them xD its totally normal.

[@Spike20101000]

WOW, The AI did better sponsor spots than the sponsors.

[@gavinhigham1478]

One of the dumbest parts of the Eufy cameras is that they stop recording if your internet goes down. Even when there's a microSD card inserted into the camera, and you've set it to record 24/7...

[@antiHUMANDesigns]

Nvidia has to be consistent and "unlaunch" the 3060 8GB, as well.

[@scarletspidernz]

You mean launch the 3070 and 3080 8GB and cut down buses

[@yellowflash511]

Nvidia doesn't care at all. They unlaunched the 4080 only because of RDNA3 not the naming

[@AllahuSnackbar270]

@@RyTrapp0 Okay, kid. Let's see what happens when it drops.

[@bosstowndynamics5488]

@@yellowflash511 They pulled the 4080 because the performance difference was larger than previous "2 GPUs one name" scandals, and the heat got a bit too much for them this time. RDNA2 (not 3, because there's no RDNA3 GPUs in the wild to benchmark against) is part of the reason the 40 series cards are absurdly power hungry (because they wanted to advance performance faster than their architecture allowed them to in order to stay ahead) but RDNA was exactly as much of a threat when they originally named it the 4080 12GB as it was when they unlaunched it.

[@yellowflash511]

@@bosstowndynamics5488 And they did the exact same with 3060 now, proving that doesn't matter. It's because threat from AMD they pulled the 4080 12gb. You might think it's RDNA2, I think its 3 whatever the case both are AMD hence the guy's statement that AMD isn't competitive is proved wrong regardless.

[@Grymyrk]

Absolutely for 99% of consumers the only way for a app notification to get an image is for the doorbell to upload it to the cloud. But as for the marketing, the image should be optional with a warning when you turn it on. And they also completely failed to secure the data, you could easily generate an encryption key to be shared between your phone and the doorbell to be used to encrypt anything uploaded from the doorbell. Also the cloud bucket should be encrypted and only accessible via an API for authorization. They can also delete the image once it's been requested or after a set amount of time. There are lots of easy things they could have done to make it secure and sounds like they've done none of it.

[@MiriaJiyuu]

This. My problem is the marketing, not the actual function that it was sending data to the cloud, I assumed it was doing this the moment I could access the camera outside my home network. They say they delete it, but no timeline to delete it's been specified. So it could be 3 weeks, 2 months, 6 months etc And everything should be encrypted. That's a no-brainer. I'm not sure why in 2022 something was unencrypted.

[@estiaanj8425]

It is true that most app notifications that include images require the image to be uploaded to the cloud. This is because the app notification needs to be sent from the cloud in order for it to be delivered to the user's device. However, there are ways to make the process more secure. For example, as you mentioned, encryption can be used to protect the data as it is being uploaded to the cloud. The cloud bucket where the data is stored can also be encrypted and made accessible only through an API with proper authorization. Additionally, the data can be deleted after it has been requested or after a certain amount of time has passed. These measures can help improve the security of the data and protect user privacy.

[@roooorey8978]

So, this is totally false advertising. Period. I didn't assume things were "uploaded" to a cloud. It is bought for its "local storage"... they should pay a price and AT MINIMUM have encryption.

[@ShampooTime]

I haven't noticed Luke this super excited about something in a while, think I should have a play with AI chat too...thanks Luke :)

[@phoenixflame9157]

Thus far every purchasing decision I have made for LTT store has been because I want to support labs and your ideas related to it. Do I strictly *need* a new water bottle? No, but I can use it and I really want to support that venture. I'm incredibly excited to see where this goes.

[@DantalionNl]

The tokenizer is not a limit imposed by the developers to throttle users, it is fundamental to the model and the number of inputs it has at a specific layer (the layer that takes and converts the tokens). It is in the same way that AI upscaling models for images are 'hard wired' for specific resolutions at certain layers.

[@atienzo98]

@@DantalionNl i read it. With it being able to remember what you wrote previously, can you get around the token limit by breaking down tasks into doable segments? It can't write an entire essay but could I ask it for individual paragraphs that make up a singular essay?

[@DantalionNl]

@@atienzo98 Depends on how 'deep' its memory is to keep track of the current context. Likely its much smaller then the input size for the tokens which would make the answer: unlikely. If the context size is significantly smaller then the model will have a hard time keeping track of what it already wrote / still should address. The effect you will most likely see is that it will start to repeat itself. Several ways around this, you can try to condense its current state and use that as input for the next step. Or you could even ask it to summarize its finding at the end so it does it for you. Interesting design constraint / decision about GPT is that the input prompt and output both are tied to the same sequence of tokens (of limited size). So the longer your input is the less output you will get.

[@liaminwales]

13:40 Linus tapes over LED's, what a hero we need more people to tape over LED's.

[@JonReid01]

There are dozens of us

[@Lem_On_Lime]

*raises hand* Also, why can't there be more red ones? Doesn't effect night vision, and doesn't effect circadian rhythms...

[@moonrazk]

I used black nail polish on the blue LED on my 2.5" HDD case because I'm using it for music storage on my notebook, and when I actually have it plugged in and am playing music it keeps blinking, annoying as hell.

[@willardthegrey]

One of us... one of us...

[@liaminwales]

@@JonReid01 lol half the team has already made a comment in support.

[@Siphonife]

Yes. I would love to see open source drops coordinated. We need support for alternative options vs the adobe suite monopolies of the world.

[@thomasafine]

The solution to the content forgery problem is for cameras to embed a unique private key in every camera in a tamper-proof way, and publish the associated public keys, and add fingerprints to photos and videos that are produced by the device. Edited video would still exist, but there would be underlying raw images/footage that could be authenticated when needed.

[@clayaderhold]

seeing linus genuinely impressed is something to behold

[@danholli123]

I love the sponsorship, makes them more memorable AND shows off tech

[@TheInfinityMaster1]

The "OpenAI Chatbox" thing is *very addictive* and can make people (especially those who "Google" things a lot) very *reliant* on it. I can totally see the benefits (and dangers) of such technology.

[@coherentpanda7115]

It's refreshing to search for an answer, and not have to sift through tons of SEO spam on Google. It may be a little too confident in the answers it provides, but for most searches, it is the answer you were seeking, without sifting through ad-filled garbage, and multiple wrong answers on shady websites that somehow Google deemed should be on the first page of your search

[@stereodark]

I asked the OpenAI a bunch of questions about Wordle. It didn’t know the game, I described it and it developed a strategy to solve it as fast as possibile. This is insane.

[@DriftGod400z]

"Wow, this video was such a great watch! As always, Linus and the team at LTT did an amazing job explaining [topic of the video] in a way that was easy to understand and entertaining. Keep up the awesome work, LTT!" - open ai

[@jeremyhopwood]

1:43:14 so I asked the Chatbot AI to write a KZbin comment advocating for Linus to step down as CEO “I think it's time for Linus to step down as the CEO of Linus Tech Tips. While he has done an amazing job building the company and providing valuable content to his fans, I believe it's time for him to step back and let someone else take the reins. This will allow Linus to focus on what he does best, which is creating engaging and informative content about technology. It will also give the company a fresh perspective and new leadership to take it to the next level. I have no doubt that Linus has the best interests of his company and his fans at heart, and I believe this is the right decision for everyone involved”

[@BobTheMartin]

I like how this episode is a tad longer, when WAN show ends it's always like "nooo plz more"

[@misku_]

How did you make me actually sit through the whole sponsors part and actually enjoy it?

[@MeowThingy]

You just missed the Lastpass data breach. Just got an email from them informing me of another one.

[@muizzsiddique]

I got an email from them 53 hours ago.

[@2trains182]

Imagine if we could actually trust any of these security camera companies?

[@ffsireallydontcare]

You can, but the trustworthy ones cost A LOT as does the management of the footage.

[@ac.creations]

roll your own? If you are serious about trusting your own hardware, there are ways to fully localize your infrastructure.

[@XionEternum]

@@ac.creations ^ This. Don't even trust an installer as there have been too many instances of installers being caught funneling security footage of attractive women who hire them to install security systems.

[@wpgspecb]

Blue iris and block your cameras from intenet access

[@buddt222]

I work in manufacturing and got a new job in 11/21. I’ve never worked somewhere before where if you don’t take a day off for a couple months you get approached by a manager and have them ask you if everything’s alright, and ask why you haven’t taken any time off. Since that happened I’ve taken 5 weeks vacation, and other than me being on a shitty shift (2:15-10:45) this is the best place I’ve worked. Hopefully I can get on first shift early next year when the first shift guy in my spot retires.

[@iwanskiwi]

Very nice show, thank you!

[@ThatFrench]

I was having trouble at work yesterday trying to automate something pretty simple with VBA for Excel. As I was watching this I described (in details) what I needed done and voila, the AI did it for me.

[@Kathlanus]

On the eufy stuff. Correct me if I'm wrong, but for push notifications to workeufy wouldn't have to do much.The device causing the push notification could generate an event token containing information about ip address and port that it tells the router to open for this. This can be encrypted with a key that was negotiated with the app at setup. This token (and only that) would have to be sent to eufy servers for forwarding (assuming no other means where set up, as most users wouldn't). The app would connect directly to the device using the info provided in the token. So eufy would only have to play middle man until the peer to peer connection is established. Am I missing something?

[@egutzait]

For the chatbot, the code/scripting stuff is really impressive, but for the creative aspect I think all it really shows is how we're not really all that creative, and if you speak with confidence you can convince people of a lot.

[@hjr2000]

I love the WAN show and never miss it. I am still trying to understand why this is 😆

[@alexanderdimitrov1134]

There seems to be two parts of this problem with the security being the obvious one however you can totally sue a company based on the false advertisement regardless off what they have on their license agreement.

[@NKCSS]

Cool that you guys mentioned Advent of Code. It's my favorite time of the year and I fully recommend it to everybody that loves programming 😊Been live streaming the challenges every day and have playlists from a few previous years if you're interested to see what's it all about 😊

[@speedyjago]

I'm so looking forward to labs with all of the input for the LMG channels and the comparison tool on the website should make shopping decisions way easier. Makes my owning several stealth hoodies and crewneck sweaters as well as the WAN hoodie worth it. Seriously thinking about getting the screwdriver and backpack...will have to look at my budget for my birthday present next year.

[@Paulrere]

One of the most interesting WAN show's!

[@ChodaBoyUSA]

Luke tells a long story about chatting with OpenAI. Linus engages fake-listen-to-child's-story-mode.

[@dragoneyr1632]

Linus was taking part in it though, as he actually tried some stuff

[@AegisRick]

The chatGPT discussion was truly fascinating. It's hard to imagine the possibilities. Basically, you have at your fingertips a receptive baby, that can digest and learn anything you feed it, and matures to comprehension and genius/expert level in an instant. You don't like the results, you wipe the slate clean and start over with different food. There will be a market for selling these AI configurations that produces the best results for a particular field. And that field, is endless.

[@pokenectionswithprofessorp2979]

*When addons were added to World of Warcraft, you could track information like threatmeters that the game didn't give you before. In a very real sense, that became a part of a wow-gamer's cognition and it changed people's cognition in wanting to optimize the threatmeter for DPS.* The argument that you make around 54:00 is part of a design philosophy called 'distributed cognition'. Basically, whatever you are using to ease the cognitive pressure of a task is a part of that cognition under that theory. The way we used to read was communal; You would read out loud, because the way things were written was without spaces and interpunction (to save on precious paper, for example), so the only way you could read was to read out loud. But some people had the unique gift of reading without talking. Later, reading shifted into something you did silently. It became part of the cognition; You don't hear the words, you experience the words. With how much we read and type today, we think a lot less in concrete experiences, but in more abstract experiences. The medium we use to think in our distributed cognition changes how we think and not necessarily for the worse. But when ancient Greek philosophers said that books would make us unable to think for ourselves; Though that exact thing didn't happen, books did change how we think and that way of thinking can be considered extinct now. In the same way, people would say that they dreamed in black-and-white. I used to dream in cartoons. The media we consume become a part of our cognitive vocabulary in that way. But this can also have an effect on what our goals are. Recently, a really interesting essay was released by Folding Ideas, that goes through the history of World of Warcraft, where at one point in it's history, you could play 'the wrong way' and not be considered a hindrance, because nobody had really figured the game out yet. However, people started to learn how to "instrumentalize"; To make winning optimally the game; A puzzle and a race to solve. When addons were added to World of Warcraft, you could track information that the game didn't give you before and those addons are built so you can instrumentalize more easily. Threat had been a fairly arcane concept before that pushed people into careful forms of play, but when threat was figured out and 'solved' with addons, people recognized that there was now a 'right' way to play. It simply eliminated certain forms of play. One might argue for the better. But the way people thought very much changed. The cognition was altered and off-loaded to another system. Addons became more complex, explaining exactly where to stand in the raid and for these addons to work, everyone has to follow the strategy the addons count on you to use. The cognitive load of coordinating with fellow raid members was off-loaded to a computer that can be seen as a kind of 41th player focused on just calling the shots. That is what we are going to see with OpenAI. Not only are you going to fall behind if you're not using it to speed up something you were already going to do; If you're not using it, you are not following the optimal strategy that the language model of the company has decided upon. We don't know what it's going to be, but we know this from capitalism; Operationalizing and distributing the labor is always going to improve productivity, which means that we are going to see businesses that rely on an AI like OpenAI to function as the equivalent of a raid addon. It may or may not tell people exactly where to stand and where and there are going to be edge cases where not following the AI's instructions makes you a hindrance to the rest of the team. Even if you don't know why yet. It will no longer your job to think about the where and when, just focusing on doing a hyper specialized task that we cannot even imagine now, that the AI has taken a good stab at improving productivity by a significant percentage, even if no one understands why it works so well. Which means that if you deviate, you're doing it wrong and you're rude to your colleagues for not following what the 'addon' tells you. I highly recommend the video that Folding Ideas made: kzbin.info/www/bejne/eHyzYnxtftSWj7c

[@denveradams4909]

At 65 years old, I have started my 2nd, high-end desktop PC build project. Started with the Segotep Phoenix T-1, E-ATX case. Seasonic Prime PX-1300w PSU. Gigabyte Z690 AORUS Master M/B. Intel i9-13900k CPU. 64GB of G.Skill Trident Z5 RGB 6400MHz (OC) DDR5 RAM. (3) Samsung 980Pro 2TB M.2 NVMe SSDs for storage. Gigabyte RTX-3090 AORUS Xtreme GPU. Corsair H150i Capellix 360mm AIO CPU water cooler. All case fans are Corsair RGB Elite Mag. Lev. I haven't made up my mind yet on what display I will eventually purchase.

[@EoRdE6]

did I arrive before the timestamps? Damn

[@bahamutbbob]

I bought a 1050 when my 660 died, right before the crypto boom. I could have sold it 6 months later for a profit, but didn't. I ended up giving it to my nephew to replace his igpu when I got a 1070 around when the 20 series came out. The 1050 worked great for me. The 1070 was a huge upgrade, but the 1050 was better than the 660, for sure.

[@Flash_345]

I tried out the chat bot myself too. It's amazing. I asked to write me a piece of code the calculate electrostatic interactions in the long range, used in molecular simulations. And it wrote it correctly. Blew my mind

[@thesuperenderdragon]

Great podcast, keep up the good work!

[@softxpandguest708]

I would 100% buy from an independently validated storefront. I'd pay a small premium to know I was actually getting what it said I was.

[@daflibble3202]

I look forward to a review one day of a proper local camera system that has no cloud calls proven by your labs.

[@deathcat1016]

I already used open ai to bounce ideas off of for an assignment. Super helpful. Feed it a book and a chapter. It will summarize and you can ask it questions about the content.

[@MusicForHourss]

I just used it to get inspiration for a card for my girlfriend with some roses. I love this function

[@lol007]

You got me so interested to try open chat ai. I tried it for dnd room description aid and it does wonders with a little direction on style and sentence connection flow. I just not that much in dnd but dont want to drag the game down for everybody so this is such a help for me ! Going to try it tomorrow.

[@bluebandit5586]

if linus wanted to keep labs data trustworthy, instead of just saying "trust me bro", open source any prporietary software developed for labs. As an engineer, I stick to open source as much as ppssible because I value transparency. I and other developers can go look at the code and make sure Linus is honest. So on each page, you put a github link with all the code and documentation. Boom problem sovled. If people can look at the code and make sure everything is above board, your trustworthiness is basically untouchable

[@jrp335]

How is it that people/companies have lost the shame of arguing against facts? I feel like people don’t ask for forgiveness, let alone permission, anymore after getting caught being awful.

[@BitHappyy]

"...ok so there's a couple of things that I've noticed: one is that I've been getting way more engagement on Twitter than I used to. *So either I'm like, awesomer.. or not all of it is real.* " -Linus trying his absolute hardest to not admit that Twitter might actually have gotten better

[@calloomdotcom]

listening to luke talk about AI makes me think they're gonna make great DMs one day

[@BNR_248]

When i built my first pc 12 years ago i spent 300 dollars on a i7-950 and 50 dollars on a ATI 5450 thinking i can run games at 720p. 5870(i think it was called) was miles ahead in performance for an extra 200 dollars. I was guilty for falling for the "gaming" gpu in my teeny days. Happens to the best of us. If your videos advising against such GPUs was existing back then it would have helped save me from making that financial mistake.

[@gamingmarcus]

My first PC as a kid was a prebuilt. And not knowing anything about PCs I ended up with an AMD APU. It said HD6000 whatever graphics on the box and I just didn't know any better. I think I lasted a bit over 1 year before buying a new system because obviously it's impossible to upgrade a prebuilt in any meaningful way. What a waste of money that was.

[@BNR_248]

@@gamingmarcus Ohh I completely understand, that's one of the biggest cons and is in line with what Linus was saying about it generating e-waste within a year. Its actually super common for an end user to want a gaming PC and know nothing about computers. They'd see a "cheap" pre-built and think its the same as a custom.

[@themohawkfoundation]

I asked it to write a poem to my pregnant wife… This amazing thing is what I got Roses are red, violets are blue Emily, my love, I am so grateful for you Carrying our child, so full of life Our family is growing, adding joy and strife We don't know what the future holds But together, we will face it, strong and bold Our love will grow, with each passing day As we welcome our little one in a special way Emily, you are beautiful, inside and out I am so lucky, there's no doubt I can't wait to meet our little bundle of joy Our family will be complete, and our love will never die.

[@m4nc1n1]

The delay is why I ordered the silver shaft screwdriver. I must say it greatly exceeded my expectations! Good job!

[@jamesj1523]

I would ditch Amazon if I found the same product on LTT Lab and saw real tested products that still met or exceeded standards. Just the thought of being able to buy a product based on FACTUAL data would be AMAZING! Link a video from testing on the product page if you sell it on LTT.

[@gammah]

The show was excellent!

[@CodyHobbs]

"I just watched the latest Linus Tech Tips video and was blown away by the AI chatbot technology. It was really cool to see Luke using it to write Python code for a battleship game. I never would have thought a chatbot could be so versatile and useful. I can't wait to see what other uses people come up with for this technology. Keep up the great work, Linus Tech Tips team! This comment was written with Open AI.

[@World_Theory]

2:10:57 Actually, I've done a bunch of encoding and bitrate tests with H.264 in a bunch of different resolutions. And very high bitrates won't hit your performance negatively unless it's simply taking up a noticeable percentage of your disc writing bandwidth. The thing to watch out for is very high video resolutions. Increases in video resolution can actually affect PC performance. I think this would be the same for AV1. Because I think the reason higher resolutions affect performance whereas higher bitrates don't (at sane bitrate values) is that the encoder has to first do the the work analyzing video frame differences, and only after that does it decide how much detail to describe the sequence of images with, and actually carry that out, which I suspect is a lot less resource intensive than analyzing the frames in the first place. Less important, but still important, will be the performance impact of encoding video as higher frame-rates. I expect that to be of similar impact on performance as video resolution, because it is resolution in a way, but in the time axis. Frame rate won't be terribly interesting to test though, because most everyone streams at either 60, 50, 30, or 25 FPS, and I don't expect that to change any time soon. But there might be reasons to record at higher frame rates. I would suggest testing AV1 performance impact with higher video resolutions and frame rate, to see where, if anywhere, game performance might be affected. Testing bitrate was a good idea anyway. I suspect you'll just run into encoding overload issues before you affect game performance anyway.

[@LabelsAreMeaningless]

My son always loved veggies. All I did was refer to them as mommy's special treats. Worked like a charm. Spoiling them by changing how you prepare them until the kid says ok, is NOT ever the way to go. It makes a kid who wants things their way or no way. A horrible precedent to set.

[@daniels-mo9ol]

As an Android developer it scares me that people defend the publishing of personal data in an app that is supposed to protect it locally. You can totally send push notifications without the data hitting the cloud lol. Its just bad developers without knowledge, or a malicious company.

[@ffsireallydontcare]

A push notifications local to a handheld device is one thing, the problem is when a separate device needs to send a notification to your handheld, especially a separate device behind a firewall or NAT on a different network. You have to somehow get that notification out of the private network and onto your handheld via the public Internet. This is especially difficult as often both devices, your handheld and your private device, will have private IP addresses as many telcos issue private IPs to mobiles. Because of this, any connection between these devices needs to go via an intermediate device with a public IP address. In this particular instance the private security footage should either be accessible on a cloud service but only via an authenticated session, be encrypted by a key shared by both your app and the camera, or held locally on the camera until tunneled by a private encrypted connection between the camera and the app via a cloud conduit.

[@daniels-mo9ol]

@@ffsireallydontcare you're wrong. The company have local storage which means it has some device capable of serving as either a polling server or a socket server locally. All you need is a cloud server that serves the initial handshake between clients setting things up. No actual privacy sensitive data needs to be sent to the manufacturer or third party. If you don't want the phones to be polling you use pushes only to notify the phone that the server wants a connection. That way you don't even need port forwarding.

[@ffsireallydontcare]

@@daniels-mo9ol As the mobile device is incapable of talking directly to the local storage due to network limitations (NAT, stateful firewalls etc), how do you propose to transfer the sensitive security footage without an intermediary server?

[@daniels-mo9ol]

@@ffsireallydontcare read up on websockets and TLS. If the home is behind a proxy, sure there may be some additional setup required but if you have a proxy, you already know what to do.

[@ffsireallydontcare]

@@daniels-mo9ol Dude, IT professional here of 25+ years spanning development, security, server and network. You've negated your own argument, a proxy IS an intermediary server. The only way a "local storage only" solution will work for normies is where the network gateway is aware of the security device's protocol. either through open standard, license agreement or closed ecosystem. Ubiquiti operates in this space. I *believe* (I haven't looked too closely at it) their camera devices send push notifications to Ubiquiti's cloud messaging service, the app receives the notification and then the app knows how to talk to the local network's Ubiquiti security gateway (firewall/NAT network gateway) for access to the privately stored footage. Without this directly Internet connected device being part of the handshake, the only way normies will be able to easily set up this sort of product is with some sort of cloud intermediary.

[@pr1vatepiles1]

That chatbot is amazing. I perked up when they started mentioning stuff about writing code. Please do an in depth video on this!

[@fortdriver]

11:42 pm here, wearing my air pods and that knock scared the 💩 out of me lol