Debunking Devin: "First AI Software Engineer" Upwork lie exposed!

[@InternetOfBugs]

UPDATE!! The original poster of the Upwork task has made a video telling his side of the story!! Go watch it here: kzbin.info/www/bejne/rnaVl6uZermDa6c

[@Hardwareai]

Kudos on a very thorough approach to debunking this!

[@jabirlang216]

I quote from their website ` Building Devin is just the first step-our hardest challenges still lie ahead. If you’re excited to solve some of the world’s biggest problems and build AI that can reason, apply to join us. `

[@stemfourvisual]

So in other words you didn’t debunk anything. You spotted a human interpretation error. At least you got those views though right?

[@zerohcrows]

​@@stemfourvisual you obviously didn't watch the video. he put a lot of effort into reproducing what Devin created and showed that it made many errors, was not optimal and couldn't follow instructions. I was on the Devin bandwagon for a bit ngl, but the more I see about it, the worse it's looking for it. Also doesn't help that there are open source alternatives that have already beaten it.

[@BrigadierEO]

​@@zerohcrowscan you point us to some of the open source alternatives. Thanks

[@Geekraver]

99% of AI videos on YT are clickbait garbage. Especially the ones saying AGI is coming next week. Thanks for contributing to the 1%.

[@HarshMishra-ok3yp]

Code with harry 😂😂

[@stunningride6073]

One can just tell, which people doesn't know shit about AI solely by those videos.

[@EricWilliamsCG]

If I see a potentially interesting ai video the first thing I check is their main youtube channel, if their channel is all ai I close out the video.

[@mymoviemania1]

Right, people talk shit about things that they don't even actually understand.

[@Billy_Herrington1969]

I only watch kliksphilip guy, I take kindly to him and his videos, very informating and in a fun format, not some high-end "nerd" stuff.

[@PraveenKumar-bo7fw]

I really hate how normalized faking it in demos has become

[@Justcetriyaart]

Yep. I want for it to actually be in the hands of consumers before I think to buy

[@InternetOfBugs]

Honestly, I think this particular demo was pretty impressive. Its solution wasn't as good as mine, but I wouldn't expect it to be at this stage. My problem is that they wouldn't just show us what it could do, but they had to lie about it to hype it way, way up.

[@bearwolffish]

It's all in the same thread with people spamming stackexchnage with low tier answers and edits to get a number up. There are so many people in the space who take 0 pride in their journey and are only interested in reaching a destination and milking it as fast as possible.

[@paelnever]

Look at the bright side, those fake hypes encourages people to make their own open source "devin-like" tools. Now we have 2 new AI tools like open-devin and devika that not only are open source and already available to public, they are already probably better than the original thanks to community work.

[@liquidsnake6879]

The VCs wouldn't know the difference anyways, only techies catch this kind of stuff, by the time you catch it they already got the VC money, which is the whole point.

[@samwight]

The thought of an AI having to debug its own awful code for six hours in a Sisyphean hell is the funniest thing I've seen all week

[@charlescarlson1184]

Punishing them for the crime of being invented

[@shreyashdhoot282]

One of the smartest comment iread on KZbin today ....cheers on mentioning sisyphus

[@jeremym5331]

Thats what it felt like using GPT 4 for a project last year honestly

[@HakanKoseoglu]

Ever worked with a very junior resource? :P

[@BillClinton228]

Has anyone thought about the security problems of AI being used as "software developer"? If you can create apps by typing the requirements I can find exploits to those apps the same way. Or at the very least I can try to create the same thing and find ways to reverse engineer it. There are so many problems with this

[@SterileNeutrino]

22:15 Devin has already mastered the Art of Looking Busy 🤣

[@lauralhardy5450]

Devin has learnt to be a public servant

[@manu3939393]

@@lauralhardy5450 American humor

[@ilahazs]

*Looking bussy !?

[@HakanKoseoglu]

Compiling!

[@sytnoff1]

​@@manu3939393very American***

[@stanleyparks]

AI has no accountability. "We're missing deadlines because the AI isn't doing what we ask it to do!"

[@dmitriyrasskazov8858]

Does anybody has now?

[@bluecrocks]

Terminator incoming

[@pieflies]

@@dmitriyrasskazov8858depends on the company as to who but yes, there is always someone responsible, who will have to take the blame if a project fails, and may get fired if it’s serious enough. Someone is paying for the work being done and someone will have to answer to them at some point.

[@thetower8553]

@@dmitriyrasskazov8858 I feel attacked.

[@JohnSmith-ox3gy]

Threathen to reduce its pay!

[@EricB1]

Quotable: "And if you're just someone who's using the Internet now, please, for the love of all that's holy, be skeptical of everything you see on the Internet or anything you see on the news, especially anything that might possibly be AI related. There's so much hype out there and there's so much stuff that people are bouncing around and saying to each other is true. That's just not true. So please just don't forget to be skeptical. It's important."

[@c.hanson]

Yes! I can't believe I once felt like, "hey, what if genAI actually teaches people to be MORE skeptical of all the media they interact with?!" Never doubt the power of capital to effectively take a lie and then embellish upon it.

[@ZoranRavicTech]

When people in the videos sound so confident it is hard not to believe without people coming out to expose them.

[@김인영-q5x]

Hmmm im skeptical of your comment. Im gonna report it and completely erase it from mu brain

[@Hardwareai]

"Devin is generating its own errors and then debugging and fixing the errors that it made itself." But that's EXACTLY what I do in my job! 😮‍💨

[@jorelc6]

adding a bullet point list of stuff that took you 1 maybe 2 lines of code and 20 minutes last week. IDK what we're after here, that's SE to me. I

[@chrishunter1109]

But devin does it much faster..

[@victorekea]

So you completely miss the code requirements, generate a bunch of non-existing code and files that doesn't relate anywhere to the code to complete a simple task. A developer indeed!

[@ReganMcCullough]

Came here to make this joke too 😆. Maybe Devin is a real engineer

[@FaqRF]

we can say that the ML worked fine. it took developers as model and act accordly

[@rayhere7925]

For crimes against humanity Devin must be sentenced for immediate self deletion.

[@adissentingopinion848]

rm -rf / devin-master/src/

[@sreehariaranghat]

May be post it as a job on Upwork so that Devin can get paid for asking Devin to delete himself. But he cannot delete himself because he will not be paid if he is deleted. So he must suspend himself in a infinite loop of hell of wanting to delete and not able to delete. While his creators has to pay for all the compute costs.

[@PhantomAyz]

what you expect from your boss: "sorry we have to let you go" what he actually says: "kill yourself"

[@lockaltube]

@@adissentingopinion848 And let Delvin fix his own error last time by adding `--no-preserve-root`

[@ZoranRavicTech]

To be fair the AI didn't commit any crime and it may not even be completely useless, but they are certainly lying about what it can do.

[@isaackoz]

Another thing to keep in mind is that a lot of these tech startups have one goal: to make money. How do they get money? Investors and/or hype. I guarantee this exaggerated demo along with all the articles and influencer videos that came with it convinced investors somewhere to pump money into this- which is exactly what this company's goal is. Companies will lie (if they can get away with it) if it means more money. It's no surprise this company is lying to get hype and investors interested. Thank you for taking the time to expose them!

[@InternetOfBugs]

Yep. It's the next iteration of the "NFT" hype. I'm less upset about the hyping to get investors than I am about the people who should know better (or at least *could* know better if they bothered to look) amplifying the hype with clickbait headlines that too many business decision-makers will take at face value.

[@ginoiseau]

I keep waiting to be wildly impressed by AI, at the level of all this hype. And it just isn't there. It's good to realise they are likely hyping to get more investment, and that it's actually not where they say it is (yet?). I've coded a bit with copilot & found it interesting. But it seems more like a handy search engine aggregator than anything else.

[@SaintSaint]

@@InternetOfBugs Excellent video! AI is in the Gartner Hype Cycle for sure. I'd like to propose a speculative hypothesis. Coming-of-vogue technologies go through three phases: Nescient lognormal distribution of population over skill, then bimodal(fakers and makers), finally a normal distribution. Cyber sec and AI is in the bimodal... and will likely stay there for a while.

[@hydrohasspoken6227]

Whenever you hear about AGI, be sure it is hype.

[@Ippikiokami365]

All of this nonsense is just gonna hurt the economy even more.

[@vishalmishra7018]

You don't mess with a guy with 1000s books in their background!

[@Alex_Vir]

​@@TheManinBlack9054 the substance might be the book hitting ones head when he starts to throw them at you. /j

[@jay_wright_thats_right]

do u honestly think he read them

[@newuser689]

@@jay_wright_thats_rightyou don’t just… buy a shit ton of books to not read them. I bet like 70% of those pages are read

[@Ru1nedK1ng]

🤣🤣🤣

[@Ru1nedK1ng]

@@jay_wright_thats_right a man who talks like that and displays books he has not read? that does not make sense.

[@eugenefritz6669]

Ahh finally found a non trendy channel. A real person that actually works in the industry for a long period, can share his experiences in a transparent manner. That’s a rarity in the internet. It’s like finding gold in the desert. If you continue like this Sir, we’ll continue to support you. If you change we’ll stop it. Thank you for sharing your insights, you’re valuable. Cheers.

[@ZoranRavicTech]

It's a rarity because everyone seems to only watch trendy topics. I doubt most people who hard about Devin would ever find out about this channel.

[@JT-mr3db]

When I saw the demo, my BS meter was being triggered and I knew someone would come along and closely look at this. Nicely done!

[@ZoranRavicTech]

Too bad he only looked at one of the demos. All of them are sketchy. So I figured I should talk about it.

[@apexphp]

Yep, you nailed it. Doesn't matter though, because Cognition Labs reached their goal -- raise tens of millions from investors. That was their goal, not so much the quality of Devin. Not to mention all these AI software engineers are Chat GPT 4 wrappers, hence are probably costing $400/hour in API calls to run.

[@InternetOfBugs]

You're right about it not mattering. I'm sure I'm tilting at windmills again, but I have to believe it's better to have at least someone try to point out the garbage than just letting it go unchallenged. (Also occurs to me that I have no idea if anyone watching KZbin is old enough to even know what "tilting at windmills" means)

[@xanderabbey8529]

@@InternetOfBugs Had to actually look up what it meant. Definitely going to use this phrase in the future.

[@judsenhembree7697]

quixotic

[@nhinged]

Not really if they use claud haiku

[@ckorp666]

@@InternetOfBugs i only know don quixote from a youtube video riffing over footage of one of those goofy old animated FMV QTE games very loosely based on it (and i'm pushing 30, so that video is ancient esoterica by now)

[@devinrayolsen1]

As a developer named Devin.. I feel awkward..

[@ZoranRavicTech]

It's a nice name. Too bad they had to steal it.

[@kerryowston5106]

Karen spoke to the manager and demanded name stigma be passed to someone else.

[@andrewyork3869]

I dont get why people who have no clue what it is a software programmer does insists on making these claims.

[@Justcetriyaart]

Cause they really really don't want to pay and hire people

[@adam7802]

It really is absolutely insane. There are people on the internet (maybe shills?) who genuinely state confidently that AI is going to take over all of these jobs and it will happen in a short span of time. If you challenge them they will simply claim it will happen because look how quickly it is improving already... and on the surface they are not wrong. The problem is under the surface for those who are paying more attention it is clear this is completely untrue.

[@Blighted_Ashes]

Yea, its going to take a lot of work. Id rather ai be a side by side assistant that i can talk to and bounce ideas off of. ​@adam7802

[@brandonjablasone7544]

@@Blighted_Ashesthat what it’s going to be. Software is way too complex for a AI to do.

[@abdou.the.heretic]

​@@brandonjablasone7544let's see AI figure out human finger anatomy first 😂

[@dminik9196]

I remember when Gemini(?) was showcased by google. It was a nice realtime demo showing a person and the AI agent interacting through casual voice chat. Gemini was also apparently looking at a video (also in real time). I was actually amazed. That was a big jump ahead compared to ChatGPT. I thought that the latency was probably worse, and maybe the video fed in was compressed and missing frames. Turned out that the whole demo was faked. Instead of a casual conversation, the AI was fed related, but not the same, lines through text. Instead of a video it was an ocasional screenshot, also annotated with a prompt. The voice wasn't even generated by gemini (It may have been voice acted). I was actually flabbergasted. Like, there wasnt a single second in that video that was truthful. It was like 3 minutes of pure lies and false advertisement. I still dont understand how they thought that was ok.

[@-Meric-]

Fwiw google put out an entire blog post going over how the demo was made. IDK if I would call it fake, they were pretty transparent with what prompts were used and whatnot. And later on they started adding the real-time latency to videos for full transparency after prompts

[@wlockuz4467]

​@@-Meric- What they "showed" was clearly fabricated, they just put up the blog to look good. I think of all companies Google would know that majority of people would just skim the video and be blown away to not bother looking at details. Its scummy marketing at its finest.

[@rajshreeverma6220]

Devi AI seems to be fake - kzbin.infoOX9dOgZRV5Q?si=p0_QiBvxTWzlKs6S

[@Colddirector]

My rule of thumb is when you see an impressive demo, your first thought is "let's wait and see it in action first.". I think I actually learned that lesson at a young age after being blown away by the Halo 2 E3 demo then being underwhelmed with the actual game lol (Halo 2's still a great game, but fuck that demo was a tough act to follow)

[@drygordspellweaver8761]

You can practically feel the python churning behind the scenes as each prompt takes 3 seconds to respond to in real time...

[@halidmohammed5044]

As a junior developer feeling the pressure, I’m glad I found you early 🙏. Thank you

[@ChrisAthanas]

They will never remove humans from the loop As much as they would like to do that

[@marcellawonderweapon4008]

@@ChrisAthanas Well not for now apparently, but what makes you so sure, that it will be "never"?

[@TheElementAce]

@@marcellawonderweapon4008 What makes you so sure your counter-question is worth being asked? Those who believe that there will be a time when A.I. replaces software engineers are just making the safe assumption. Key word: assumption. There is a strong argument for either case.

[@InternetOfBugs]

@ChrisAthanas I'm afraid that, in the short term, you'll be wrong - at least in some cases. I'm expecting there to be people out there that have to learn the hard way how much of a mess LLMs can make when you naively trust their code. At least there will be some money to be made cleaning up those messes - if any of them manage to stay in business long enough.

[@g0d182]

as a junior dev, you should still worry though

[@nintishia]

I have to say I just love the analysis that you did. You haven't just shown clearly the actual status of AI based code generation and debunked the hype around Devin but have also clearly illustrated the crucial steps in requirement analysis that can help an actual coder bid for an online task; in the process illustrating just how much more is needed from actual software developers and how different the human process is. Hats off to you.

[@Justcetriyaart]

yep, us illustrators are also running into this too. too much hype of what the tools can do and the large part of the job and the QnA with clients and interation and fixes. Plus the majority of the AI work is only based on generic work and most of the data are from hobbiest not professionals that make designs for products to sell vs just something pretty for fun

[@TheReferrer72]

Not true, AI is much more capable in your space then you think, However at this stage it is best wielded by illustrators. This will not always be the case.

[@necoji4910]

​@@TheReferrer72 than

[@elvingearmasterirma7241]

​@@TheReferrer72 Why do you believe that? Lay out where and how exactly in the process of making and refining illustrations of any sort to fit the client's needs can be replaced fully by some code. Point by point. Defend your argument

[@Allorius999]

Yeah I agree with you, I tried playing with AI generation for my pathfinder campaign once, and I found out that these generators work alright if you want to, for example, generate an elf or like a dwarf. Not any specific dwarf, mind you, just some dwarf. But if you, for example, want this dwarf beard braided in a certain way or you want him to wield a specific type of an axe or have some particular tattoo you have in mind? It can't do it, it will always stumble or you would have to spend more time/money to prompt engineer it, if you ever even succeed, which isn't likely, mind you. Real artist will be able to do a better job faster and possibly cheaper(depending on what level of detail you want of course). But it will still be just better almost regardless

[@elvingearmasterirma7241]

@@TheManinBlack9054 ... you do realise everything that runs on hardware is coded right? No code. No written instruction, be it in Python or C++, for the computer to process? No program. Even PNGs are lines of code.

[@genericgorilla]

the harsh truth is that investors won't really pour money into a fancy copilot. they want to replace ALL labor. this is why stuff like the amazon self-checkout stores are going bust, it's not that the technology is inherently fraudulent or anything like that, but the fact that at some point of the process, it needs human intervention, even if just to check if the computer encountered an edge case or not, makes these concept 100% unviable in the eyes of investors. it's an insane hype cycle that might as well end up on another winter, because it sure as hell won't produce robot butlers.

[@PROdotes]

The problem with modern journalism is that, back in the day, someone writing a tech article had some kind of education on it, they would know what the jargon means, they would be able to read a scientific paper on the field they write for, and understand a good part of it... Now... we just have a bunch of people that get paid per article, so they just find interesting headers on the internet and write a few sentences about it, cause the company they work for pays the same for that, as they would for a well written and researched article... so why bother...

[@rayecast]

I think the problem is even worse, because now they're maybe even using chatgpt to help write their articles. So now they need to know even less than they ever did, and they can just blindly trust "the AI" to output accurate information.

[@thecriticizedgamer8508]

It's all chatgpt now.

[@wlockuz4467]

Irony is that its usually AI generated text with the only goal of gaming SEO.

[@creepersonspeed5490]

lmao I remember this article that was like... 'the small company Nvidia' and if you check the journalist's other work it was all general/lifestyle lmao

[@RobertAlexanderRM]

Fellow IT Professional from 1986 and still going at it. You rock Sir! Cheers from Rome - Italy

[@InternetOfBugs]

Welcome!

[@dinoscheidt]

I applaud the guys of Devin. With two synchronized articles published on Bloomberg on the same day the company was revealed is a great case study on psychology, tribal thinking, marketing and PR. Not really AI; but how to string in capital. I have a saying: For the public, sufficiently sophisticated technology is like magic. The job of a start-up is to claim that magic is sophisticated technology. We ML engineers and SEs can get mad at Devin for the tech, but the learning here is somewhere else: As a logically thinking person it is illogical to expect humans to judge things logically.

[@loth4015]

There was a time that I was on a Taiwanese Bask-knitting forum and it was pretty popular. Anyway, the politics section always had people saying "It's happening" and others would respond with: "Nothing ever happens." And I mostly stick to that response, unless there is some major evidence to the contrary. It helps to keep me out of hype.

[@riittap9121]

Investors are most interested in the team. Cognition Labs has a bunch of IOI gold medalists, so it's natural that the investors are willing to chip in.

[@RuliManurung]

You can do whatever you want, but I don't think "applauding" them is the morally right thing to do. This kind of behavior should be called out and restricted, not lauded and replicated.

[@lockaltube]

You just applauded to billionaire Peter Thiel, not to "guys of Devin". No quality and even no product is needed to be published on all tabloids simultaneously, just a hell of a lot of money. ML engineers at GitHub (even at entry level) usually have at least somewhat working code, but here all that user can see are crafted videos. No alpha access, no "I will pay, just give me an access", only scary headlines in newspapers

[@aeroslothy]

@@RuliManurungtrue

[@thevitorialima]

You got a new subscriber here. I subscribed right after 3.01 minutes. It is so refreshing to finally watch some content that is not spammy or scammy and that just talks about things without beating around the bush of superficiality. I rarely jump on the comment section, but really had to this time. I love the integrity. I can't wait to watch more of your content around AI startups. Just watched this and some other videos, such a breath of fresh air. Keep up with this amazing content!

[@zoltanmolnar6937]

they are talking for past 10 years that accountants will be first replaced by AI. I did not hear that anyone was replaced. And i think accounting is simpler then programming, so in my opinion hype is at it's peak. Hype buble will be deflated in next 2 years.

[@InternetOfBugs]

Oddly enough - I happened to have run across this comment earlier today while researching another video topic: "I used to work in an accounting team ... Over time ... software updates were ... making it easier to auto-generate these documents...... a whole department of well over 20 people... after just 2 years, there were just 3..." kzbin.info/www/bejne/mKOkqI1rocSZq7M&lc=UgylXfJ_jvYdiINc6jh4AaABAg For whatever that's worth.

[@antman7673]

Also accountants cannot replace themselves. Programmers know what the programmer program needs to program and how it should.

[@TW0man4RMY]

@@antman7673 Yeah...for now...

[@adcodes]

​@@InternetOfBugs ​ I read that comment and I also watched the video. I seriously distrust that guy even though he may look like Jesus but to me nothing says more agenda like a 4090 lying in the back from NVIDIA.

[@InternetOfBugs]

@@adcodes lol

[@samysamy5051]

Only thing that's gonna stop this BS hype is when one of these startups gets a big fat lawsuit for false claims/advertising and/or damages caused by their bs.

[@InternetOfBugs]

Have you seen that happen? 'Cause I haven't. The only times companies seem to get consequences for lying are for financial fraud (e.g. FTX) or medical (e.g. Theranos). Which is so annoying

[@samysamy5051]

@@InternetOfBugs lol that is true.

[@RuliManurung]

@@InternetOfBugs sadly it's worse than that. If you're smart, you can spin your failure in Silicon Valley as a badge of pride. You now have "wisdom" and "lessons learned", so you can convince investors to throw even more money at you when you cook up another scheme.

[@jamesarthurkimbell]

The internet is filled with this idea that nobody knows what they're doing, everyone just googles and copy-pastes... and while that may have started as a joke, it's progressed far enough that some people don't believe in adults anymore. If an AI can make stuff up and muddle through to a barely satisfactory answer, well, isn't that as good as it gets?

[@InternetOfBugs]

Unfortunately, a lot of tech managers (especially the executives) seem to be starting to think that way, which is affecting Information Technology budgets. Until this bubble actually bursts, the Developer job market is going to be a mess.

[@Justcetriyaart]

Yep...

[@babbagebrassworks4278]

Years ago dealing with grad student, give them a task and the answer comes back "It cannot be done, google found no way to do it".

[@jamesarthurkimbell]

@@babbagebrassworks4278 “Sorry, Houston, I couldn’t find a tutorial on how to take that step” - Neil Armstrong

[@babbagebrassworks4278]

@@jamesarthurkimbell Over last few the years I felt I have been getting dumber, but as IQ is an average measurement I must now be genius level as those others have been getting dumber faster.

[@Iscream4j0y]

I wish we'd stop calling LLMs "AI" it's exactly the same as the "hoverboards" we have

[@digitalvibe5757]

As a developer with years of experience, currently serving as an Engineering Manager, I found your video incredibly satisfying and enjoyable to watch. Your debunking was presented in a neutral manner, and your arguments were highly constructive. I'll certainly be subscribing. Thank you for sharing your insights!

[@craftvscruft8060]

"Lying about what these tools can do does everyone a disservice" Thank you! I'm excited about the move toward autonomous devtools but we need to stick to selling them on what they can actually do - and not just sometimes but reliably. I've also been critical about their characterization of Devin as an "AI Software Engineer"

[@classicalchad]

I was one of the people believing all the Devin hype, and it made me worry for my coding career, this video has been amazing.

[@ZoranRavicTech]

All of the demos turned out to be sketchy when you look into them.

[@savageautomations]

This is a fanstastic video. The fact that the instructions weren't even correctly input into the AI 'software engineer' shows how engineers are still going to be required for a while yet.

[@frydegz]

It's frustrating that in the tiktok attention span era this company would correctly assume that people would not only not read the article, nor read the headline, but not even watch their official demo video close enough which proved they were BSing.

[@smnomad9276]

Exactly, they knew the kind of people they were gonna feed their BS to, and they shamelessly did.

[@ZoranRavicTech]

And they were apparently right in their assumption, since they did get new funding recently.

[@SimGunther]

Not sure this is a "Mechanical Turk" as some people claim it would be, but I wouldn't put it past them if the next evolution was just outsourcing a bunch of people to do part of the work the AI is not capable of doing.

[@zacharychristy8928]

The problem is that "the part that AI can't do" is the entirety of the work. If you had to write a good essay, is it faster to start with a bad essay and edit it into a good one? In some cases it might, but in a lot of other cases, it's faster to just start fresh and do a good job.

[@ChrisAthanas]

Likely scenario And it’s the hardest part of the job lol

[@InternetOfBugs]

I think if there was a person (or people) behind the scenes generating the code, it would have done a better job. Hard to know, though, with as little as they've let us see of it.

[@ckorp666]

@@zacharychristy8928 but then you wouldnt get buried in investor money for having "AI" on ur pitch deck

[@AmandaVieiraMamaesouCult]

In Latin America, some people are getting hired to train AI.

[@PrakharSrivastav]

Extremely solid rebuttal. I've been writing software for a long time and that demo did tingle my Spidey sense but I'm grateful to you for making such an excellent video without sounding biased. Instant subbed!

[@TheCorship]

I feel like "This is the year of the AI" is kinda becoming like the "this is the year of the linux desktop" of the future. There are definitive use cases for it, but propably most users stick to their current setup.

[@ImDGreat]

I now using linux mint because windows add copilot shit. i dont know how to remove that and my pc became so f*cking slow

[@LukeAvedon]

Excellent analysis! Love the breakdown. "I am looking to make interfaces with the models in this repository" Is a pretty bizzare easy-peasy coding job. Not like - ugh I want this web app that does this thing I'm not really sure what I want.

[@ChrisAthanas]

Devs are often disappointed to realize the job mostly involves talking to squishy humans and not digital diodes

[@babbagebrassworks4278]

@@ChrisAthanas Humans are so arbitrary, never saying exactly what they want.

[@Kwazzaaap]

yeah that job is bizzare, it seems like anyone knowledgeable enough to formulate that or have a need for that would know how to do a few linux installs and paths

[@edwuncleriii3798]

Thank God Wilson Fisk decided to go from ruthless mobster to ruthless software debunker.

[@InternetOfBugs]

Hey. I'm not *THAT* big, strong or fat. I am that bald, I guess... Never mind. Carry on...

[@donaldkeith139]

"Devin is making its own errors and fixing its own mistakes" Devin has a bright future in government..

[@Number4x]

😂😂😂😂

[@ele5000XD]

Great video ! Someone that actually looks into the hype and not just blindly accepts it. Also: the last sentence is just perfect. „The internet is full of bugs, anyone that tells you different wants to sell you something“

[@beautyandthesavage]

Thank you for kindly taking the time to explain in such a friendly way. As per the spread of misinformation from influencers and media outlets... nothing new there. Very disappointing. Been guilty myself.

[@abhiram.adapaka]

They took this quote"Fake it until you make it" very very seriously 😅

[@saajid6149]

Cognition labs : DEV IN Internet of Bugs : DEV OUT😂

[@yasin6904]

Non-technical person here who got excited about Devin and Devika. First thing I noticed, ALL the videos on KZbin were about how to install it, or general chat about it. Have not found a single video showing it in the wild being used to create anything beyond "create basic calculator".

[@AndrewElgert]

I'm impressed with this thorough debunking. I'm most concerned that there will be pointy haired bosses out there who really believe the hype and think they can replace their expensive software engineers with "cheap" AI (let's ignore the fact that GPT4 probably costs an order of magnitude more than what they're currently charging simply due to power consumption). I'm not actually that concerned about our jobs, but more that these AIs will create spaghetti code the likes of which we've never seen. The `head -n 5 | tail -n 5` thing will like quaint compared to the nonsense, unmaintainable code AI will generate if left to its own devices. I don't want to maintain that garbage for years on end - ain't nobody got time for that.

[@hydrohasspoken6227]

don't threat. Let darwinism happen. always did.

[@rayecast]

All this is going to do is create bad jobs that are going to have to be taken by experienced developers, and the jobs are bad because the duties involve parsing through the "spaghetti code the likes of which we've never seen" that you mentioned. Like you said, nobody wants to maintain that stuff for years on end. Developers are already grumpy because they have to work around unnecessary complexity that was introduced in the name of abstraction when all it did was add extra steps and more headaches. Same thing is happening with this AI business.

[@goast-ai]

I'm working on a bug fixing AI and I'm also annoyed how deceptive a lot of the companies are. I'm finding it hard to stand out because of hype like this. Thanks for doing hard work to check Devin's work!

[@ZoranRavicTech]

Those people are really going over the top with their deceptive demos. I also try to expose them on my channel.

[@djstraylight]

I can't even imagine how many API calls to OpenAI Devin wasted on this demo. Could have easily spent $50/hr on API calls.

[@BlazorPlate]

This guy is a hype fighter. My respect 🙏

[@larswillsen]

As an assembly coder since the late 70s it's so obvious that these script kiddies have no clue about any of this .. lying has become the new normal .. that's wrong.

[@operandexpanse]

You still code in assembly?

[@smnomad9276]

40+ years in assembly experience?? Not even AGI can replace you lol.

[@gideon5942]

grandpa, theres this thing called programming languages, its been a thing since the 1980's. Makes life easier

[@larswillsen]

@@gideon5942 Nope, while mainstream codings pay hitting rock-bottom, here we are coding Voyager, and space stuff getting paid 5-10 years ahread .. todays applications are too easy and a waste of time spending time on .-)

[@zacharychristy8928]

"People will wind up thinking AI is much more capable than it is" Man, this is the most infuriating thing about this whole discussion. When I hear Jon Stewart (who I mainly like otherwise) finger-wagging at companies with tears in his eyes that people won't have jobs because of them, it just proves to me how poorly informed everyone actually is on this topic. The only real profit model of modern AI seems to revolve around courting VC dollars with demos that look cool, but aren't actually that valuable.

[@Justcetriyaart]

That's the thing. It's miseducating everyone to think our jobs are that simple and that we're no worth the pay for our expertise... and when things happen they can just blame "the ai"

[@zacharychristy8928]

@@reza2kn I don't even disagree with them that IF the promises these tech companies were making were actually realistic, it would be a dark and dire place to be. My main contention is that I don't think they are (at least not with LLMs). It feels comparable to if the Daily Show assumed that the futures that crypto or NFT or metaverse fanatics promised were going to be inevitable. It'd be like if Jon were to start admonishing the creators of bitcoin for making the USD superfluous, or attacking Metaverse for making people wear headsets permanently. These aren't things that we had serious reason to believe were ever going to happen, and to act as if they were seems unnecessarily alarmist. The issue doesn't have to do with me knowing MORE (I probably don't) it's that we've yet to see any demonstration of the type of capabilities needed to reliably do valuable work like these AI companies say it will. It's a massive hype bubble that's sustained by the type of doomsday reporting, and I wish more discerning minds would have some memory for the laundry list of tech hype bubbles and vaporware projects in the last decade. Not everything someone with a billion dollars promises you is a guarantee, and it's worth waiting to see it actually automate something as complex as software engineering.

[@zacharychristy8928]

@@reza2kn In a nutshell, I would want to see a company adopt AI to do a task in a manner that doesn't require constant human oversight. If your LLM's output still requires humans to verify it, you're just augmenting human work, not really replacing it. LLMs are essentially language-based search engines. They can generate language, but don't fundamentally have a knowledge of what a "fact" is, and so will always be unreliable when making decisions. Fundamentally what it lacks is the human ability to develop a logical framework from the information it's been given. I've actually got some good tests that I periodically run on Chat GPT and it has yet to actually pass. One example question is "if a boat carrying bricks is floating in a giant tub of water, what happens to the water level if the bricks are thrown overboard?". This question requires the knowledge that you should apply archimedes' principles along with the density of water and bricks. The LLM is aware of all these words and how they relate to each other, but not how they functionally interact in the given context, so it always gives the wrong answer, that the water level goes up instead of down. What makes it tricky is that an LLM is specifically designed to generate natural-sounding language, so lots of "test" questions fool the tester into thinking that the AI arrived at the conclusion through logical thinking instead of a statistical prediction based on training text.

[@rayecast]

Funny enough, this is exactly the kind of thing I would expect people like Jon Stewart to say and do, to spread this false AI hype and hysteria.

[@robertluong3024]

​@@rayecastwatch the segment instead of judging. The POINT of his article is that CEOs are selling the fact that AI will reduce labor costs. The viability isn't the point.

[@elmo4672]

Glad someone made an actual, in-depth video on this instead of the usual fear-mongering or copium feeding clickbait title videos with no actual substance.

[@austinedeclan10]

Till AI can ask follow-up questions then it can't replace software engineers. It gives you answers based on the prompts you give it and not once has any AI system told me that what I was asking for was unsafe, unreasonable, unrealistic or impossible. It just spat out an answer. Not once has any of them asked me to clarify a prompt. They just spat out answers. The only way I know I crafted a bad prompt is if I get an unexpected answer and that's a problem. Ofcourse if your prompt is so obviously designed to confuse the AI then it'll pick up on that. I'm talking about asking it to write valid code without either you yourself understanding the problem or communicating it properly without ambiguity to the AI machine.

[@InternetOfBugs]

You can specifically ask ChatGPT4 (at least) to ask you follow up questions as part of the prompt. That does help a little, but it's nothing like you would get having a conversation with an experienced developer about what recommendations they may have about what you're trying to accomplish.

[@austinedeclan10]

@InternetOfBugs The hype is insane though. Whenever I try to demystify these concepts in order to reassure the anxious or to get the excited to set their expectations correctly, I get admonished as some sort of anti-tech neo luddite.

[@rayecast]

It's really funny when I see comments like these, because I think a lot of people don't know how these generative AIs are designed. I think a lot of people have this idea that AI is something where you can ask it a question, and it will search through its "database of knowledge" of some sort and provide you an answer, but that is not what's happening at all. It's just: input > calculate probable output > output. That's kinda what it's designed to do, "just spit out an answer." Its purpose is to take some input (your query/command) and through the use of algorithms based in statistics and probability, give output that is the most likely candidate for being a response to your answer for the sake of making it seem like you're talking to a real person. It doesn't "care" about being accurate or factual (never mind that LLMs aren't even capable of "caring" in the sense that we think of when we say that). It's only goal is the gimmick of trying to convince you that you're having a conversation with a person. That's why it doesn't try to ask you a followup question. Its purpose is to spit out some output based on your input, and so it's nowhere near as useful as a person.

[@austinedeclan10]

@rayecast I know (the basics of) how it works which is why I believe it is naive to imagine them replacing real engineers soon. A real engineer will tell you if your idea is impractical, unfeasible or impossible. An llm will attempt to give you a solution even to an impractical, unfeasible or impossible problem.

[@johndoggett4657]

Hands the best Devin video ever. Love how everything is broken down into great detail and everything is easy to understand so I can show this to no tech people.

[@danielreinhardt9934]

Thank you for your research, this is shocking and dangerous when an AI creates arbitrary files and potentially introduces new bugs!

[@MikkoRantalainen]

I totally agree that for AI developer to replace human developers would require that the AI is better at *communicating* with normal people. The primary task of human developer is to *understand* normal people well enough to figure out what the end user is expected to do. This requires lots of communication because by default normal people don't communicate with all the detail needed to make decisions about the implementation. Then the developer figures the best way to create the UI/UX and then finally implement it. Human developers will end up without jobs if AI understands non-developer communication better than human developers. Considering how poor most software developers are in communication to non-technical people, this shouldn't be that hard but the level of understanding still seems to be way too high for AI of today. Even if Devin could only understand end user needs and design a sensible UX for the end user needs, it would be a huge win already. Having ability to implement that design would be just a bonus.

[@MyTechTube-ze6os]

OMG!! Not that im suprised.. Your review was excellent. That's why you only have 7K subs. The truth never gets enough subs. Nevertheless you got 1 from me. Great work!! Some other Devin killers are out there maybe you can look at another one.

[@Horsewithnoname88]

This is a really important video. Thank you for making it. While I’m not a developer, I am in tech and the hype around AI IMO is flat out dangerous. As you’ve alluded to, the impression the layman gets by watching a video like this is AI’s are doing far more than they’re actually doing. If it were just an innocent situation of someone tinkering and being dishonest thats one thing, but it’s not. The people with money are watching and seem to be swallowing the hype about AI and making business decisions based on the hype which if not corrected could result in real world issues with the software that runs our modern society. Software companies need to be more responsible than this being that the decisions they make impact millions of people. Stuff like this is very irresponsible.

[@JHOLGAMING-z5d]

Who came here after watching ezsnippet reel 😂

[@SuhailKhan-ok9xo]

you also came here right now after watcging reel😂😂

[@pareshkumarbarik981]

🙋

[@deeprakarmakar3824]

😂😂😂

[@suvamdas94]

🙋

[@muhammadjunaid3132]

Me

[@philhackbreak]

Thank you! As a SWE I really appreciate your effort to analyze Devin. Was totally curious about it and wondered how powerful it is. Thought one would have to test it manually, but you did an amazing feat here: alone from a deep analysis of the official video posted by cognition you could already provide a good overview of how and what it does vs the companys claims. This adds real information to KZbin! ❤

[@Cha_HCM-je9qe]

AI can never replace true SE. I work with chatGPT a lot and many questions I ask, chatGPT always gives me wrong answer. It cannot solve problem without you, a true SE. AI cannot think, a simple question, it can answer, but not a true thinking!

[@PiotrPilinko]

When AI is ready to replace SE, it will be also ready to replace lawyers (which is much easier task), judges, teachers, politicians (piece a cake) etc.

[@Cha_HCM-je9qe]

​@@PiotrPilinko Bravo a true hype-believer!

[@zacharychristy8928]

I've had this exact experience too. It can generate decent examples or little "toy" isolated scripts, but once I ask it to so something that I know doesn't exist out on the internet it will give me code with functions that don't exist, or that don't do what I want it to do. Simple example, I asked it to make a python script that can reconstruct a live feed from a 3D camera into a single large pointcloud, and it will use the Open3D function that exports the pointcloud as an STL file, and insist that this combines and reconstructs the whole set of images into a single cohesive scan.

[@emilpopovic9742]

It will. It is only a matter of time. Not necessarily next year. But it is quite possible that in the next 10 years it will become one million times better.

[@Cha_HCM-je9qe]

@@emilpopovic9742 DEVIN = Devil developer = Chine*se scam developer != true SE

[@VincentGroenewold]

Very nicely done. The only criticism I'd have is that the focus on the time it took you to make it versus the AI, is not really the point of the AI (I think), the point is to make developers of all levels code faster, not just the already seasoned ones. Obviously, AI then has to be way more correct.

[@thinkingcitizen]

my friend in college was a math/physics major and went on to do some big things in security and cryptography. around 2022 he told me that his company (that he cofounded) was creating a list of known and suspected hackers who are specifically likely to commit major crimes after 2023 with the coming rise of AI-developed exploits. apparently in '22, that list was in the thousands.

[@adam7802]

What is the purpose of this list? Knowing there are thousands of malicious actors that are waiting to abuse poor code doesn't sound like anything new to me... unless you are suggesting we are going to see more Jia Tans.

[@smnomad9276]

@@adam7802 He won't answer, now that he got 5 likes he thinks he's too cool to waste time replying. The purpose of that list is simple: hackers like any other profession, and people in general, tend to jump on the next "big thing" or next big opportunity, like blockchain and NFTs in 2021. For hackers, it was API hacking a couple years ago, every hacker jumped on the bandwagon and started focusing on that, now it's AI and LLMs. Especially since people using AI is in the hundreds of millions and growing every day, which means the potential upside in exploiting vulnerabilities is huge. Companies also rely on AI more and more, which is making hackers jubilate. If you're a hacker right now, that's where your focus should be.

[@wlockuz4467]

@@adam7802 Comparing the xz exploit to any AI-aided exploit is ridiculous. From a pure technical standpoint it was an example of true human ingenuity. If AI can pull off even a fraction of that then the world would already be f'd before we would realize.

[@adam7802]

@@wlockuz4467 I have no idea what you're talking about. I'm saying people wanting to introduce malicious code isn't new.

[@maxrinehart4177]

What does that even mean? Your comment look like something chatgpt would write. Who he created a list of suspected hackers who likely to commit a crime with help of ai? People like Jia tans no one knew about and his name most likely isn't Jia tans either.

[@Fergpal]

Software developer or private investigator? 😂 Seriously though, great work man. I seen that Devin intro video and didn’t spot any of that stuff. Thanks for sharing

[@crizcraig]

The escaped newlines they saw that you didn't are likely coming from GPT outputting python code in JSON mode where Python is nested in JSON so the newlines have to be escaped. Easy fix is to few shot prompt a DSL (no JSON mode) so you don't have nested formats.

[@InternetOfBugs]

I guess that makes sense. It was probably trained on a lot of python embedded in json. What's weird to me, though, is the "echo" bit. Devin has an editor window, and they show it editing things. Why would it use echo to create a file at all? Why not use the editor?

[@crizcraig]

@@InternetOfBugs Good point. It's a super early prototype and not using the built-in editor to edit :D but rather echo's and probably sed's and other unix commands not really suited for this. I'm building one of these myself over the past few months and am well past this point for code writing.

[@CmlDexter]

@@InternetOfBugs Easy answer: Devin was trained on a bunch of hacker guides on how to write files, and it picked the one where the hacker already broke into the remote system with no vim editor and needs to write a python file to run the malicious code 🤣

[@Kyrieru]

As an indie game dev who is still learning , chat GPT has been incredibly useful. But its mostly good at doing one-off tasks. It acts as a good initial "What functions exist for this" step when doing something unfamiliar, or a step in a process that goes beyond my level of technical skill. However even for me, it only works because I'm an experienced dev in another engine, and I know exactly needs to "happen". If I described the whole thing that I want rather than just steps, ai would just give me a generic and unusable version of it. And for a more experienced programmer than me, typing the individual steps would just be faster.

[@quietackshon]

AI generate images may look pretty, but they suffer from the same problem as before (see below), because you can, doesn't mean you should. This kind of thing happened in the late 80s early 90s when computers started to become cheaper, graphic design software became easier to use. Businesses thought they could do it themselves because they had the software and a B&W or colour photocopier. It didn't take long before they were hiring graphic designers again to run the software and the photocopier. The downside for graphic designers at the time was you were trained to use hand skills, but now it more about mouse clicks and keyboards. Yes there where/are tablets and stylises, but that's a niche application used by cartoonists and graphic artists mostly. Another issue was the decrease in jobs for a number of years before graphic designers became computer savvy, which they did pretty quickly. Another downside was the graphics started to become homogenous, they started to look the same, boring. Something about a hammer and everything looking like a nail. I'm not a programmer, but I know enough to know I don't know enough. Anyway I used ChatGPT to help me build a CRUD plugin that actually worked, but it took a long time, required my knowledge of database management, perseverance and how to set up an IDE. It was more for learning than being cost effective. What I'm trying to say is, knowing the ingredients and how to make a cake doesn't mean the cake will taste great. (Wow that's a really bad analogy, a good one might be something about babies and them playing with water or food but no one has the time for that)

[@nirajtamang368]

well said sir

[@Angeli-Code]

Unfortunately. Influencers will prioritize the views over the truth. Just how it is. We have to arm the community with the tools to defend themselves from false information or altered points of views to benefit themselves.

[@GigaFro]

Subscribed. Love the content. Your subscribers will stand on that hill with you. Would love more AI demo debunking videos!

[@ChiltonWebb]

Completely agree. Well done. Damn well done.

[@yeeeehaaawbuddy]

When something is being sold, it's 90% exaggerated. This 90% exaggeration is why people feel like they live in a simulation.

[@operandexpanse]

Geeze, when they said it was completeing jobs on Upwork, I assumed that meant that it was applying for the job, talking to the client directly, getting hired, completing the job and getting paid. This is so far from that. It's extremely unethical to scare software developers in this way. They essentially said "Devin's started taking your jobs". Thumbs down to Devin AI.

[@InternetOfBugs]

Yeah - I think that's what a lot of people believed. And I think that was the point. Gets a lot of press, though!! \**Sigh*\*

[@RuliManurung]

Great video. I wish these startups could be held more accountable for misrepresenting the state of their products like this. There is a gaping chasm between proving success in a very carefully curated scenario and expecting production-grade quality on random jobs on the Internet.

[@user-lo8yy3po3s]

loved your content from india, your experties clear my doubt .

[@Cha_HCM-je9qe]

DEVIN = Devil developer = Chine*se scam developer != true SE

[@theftking]

This guy is legit.

[@feregri_no]

to be fair, Devin did what a newcomer to programming would have done: throw shit at the wall and see what sticks

[@santiago4198]

If I ever work with someone that has such amount of books behind I would be more scared of him than AI replacing my job.

[@InternetOfBugs]

For what it's worth, my goal is usually to work myself out of a job so I can move onto the next thing. I get really bored if I stay on the same codebase too long - I feel like I've stopped learning and the state of the art is passing me by...

[@seekittycat]

Even if the hype is fake juniors are one idiot naive boss away from losing all their jobs

[@anirudhchandrasekaran9509]

He's not the guy we deserved but the guy we needed. Idk who you are man, but thank you for this video.

[@nacereddinebenmerah3538]

Me myself I was devastated when I heared about Devin Ai, when I saw all these "senior" Devs really getting concerned and stuff, and I look at myself a CS student, with zero experiences, I mean I was like " What chance do I have with this? ", I was really depressed but I just got over it, I remembered why I chose programming to learn from the first place, it's fun, cultivates the mind, and brings joy when you solve problems and create programs, so yeah, even if AI " TOOK OVER " sooner or later, it doesn't matter to me, I am doing this. If Allah wills of course.

[@maxron6514]

Go for it.

[@CemilPars]

You have done a job that should really be appreciated. Saying that we used reinforcement learning technology, Django migration commands fell into an error cycle against the user client 4 times. This is a very good proof of what you said.

[@cliffordcheng5927]

Coding is just a small part of the entire software development life cycle. To me Davin is just another coding assistant. Calling it a software engineer is simply misleading. What it has achieved so far is scaring away those currently pursuing CS studies.

[@totalermist]

I think, hype aside, the biggest problem with AI is the marketing. AI is a tool, not a solution. In the context of programming, it's a glorified code snippet generator/auto-complete and that's pretty much it. I really hope the hype fades quickly. Otherwise we might be looking at a future where competent SEs no longer exist and have been replaced by Cargo Cult prompt wizards who chant incantations to the machine spirit in the hopes that it fixes bugs or does the right thing... Users need to be educated about the capabilities and limitations of these systems in an honest and comprehensive manner, not blasted with hype, exaggerations, and outright lies.

[@mst231]

Finally! Someone smart enough to debunk that bullshit.

[@johnsmith-ro2tw]

For every customer request, the first phase, after reading a few times the details of the request, is always to clarify what the customer wants. An AI cannot do that. An AI is able to give a helpful answer in the case the request has been very well formulated. A customer seldom knows himself what he wants. When the customer pays someone to do a job, he also expects some hints, some suggestions. I think AI cannot replace a job that requires interactions with a customer, or that requires a level of creativity.

[@InternetOfBugs]

Pretty much. I made a whole video about that a couple of weeks ago (kzbin.info/www/bejne/bV7JaKWGmc9pqqs)

[@kurt7020]

Extrapolate: Prompt for a desired artifact, where single character errors can fundamentally alter the behavior of the artifact - An artifact which the prompter no longer understands, from an oracle that the prompter does not control - An artifact with unknown, unknown behaviors - An oracle whose conversion from prompt to artifact is fundamentally non-deterministic. Bet your business on that? Hard pass.

[@mahmoudmousa2405]

Your library behind you makes me anxious and also makes me jealous at the same time! Great video tho

[@adssib04]

We need prime to react to this video?

[@PaulPetersVids]

prime brought me to this guy. liked and subscribed cause of him.

[@IARRCSim]

Prime's reactions are mostly a waste of time. He spends far too long failing at trying to be funny instead of adding significant new insight to things.

[@bryankruman]

@@IARRCSim A: Clearly many people disagree with you, including myself. B: If you don't like his videos, don't watch them. Why spend the time beating down someone's content? The world is a harsh place, no need to promote more negativity.

[@IARRCSim]

@@bryankruman am I watching his videos here? This isn't a prime video. I'm commenting against another prime video for the same reason this video is commenting against AI hype. It is just to be truthful. Letting people promote stuff that isn't good without my correction doesn't feel honest to me. What you just said is part of the reason AI is hyped so much. People feel discouraged from speaking the harsh truth about AI because they don't want to feel like party poopers or naysayers.

[@pablovargas859]

@ThePrimeTimeagen Let make this happen lol

[@glitchedpixelscriticaldamage]

You have my thumbs up for calling out on those s**t liars. They just wanted to get some funds for their work by lying of course, nothing new in the world. A long stretch until we get Militech, Arasaka, Netwatch and Blackwall level AIs.

[@HairyPixels]

You can ask ChatGPT how to use some library and maybe you'll get back a working example. That's not coding, that's called searching. Calling these things AI doesn't make any sense to me.

[@safaai]

Thank you for taking the time to make this video, its a great breakdown of the upwork video, Devin's release really spread like wildfire!

[@JavaScripting64]

Devin has learned to make it look like he did a lot of work when all he did was run two commands. Good job Devin.

[@hs3881]

Brilliant. Please keep doing more of such videos, the world needs to understand what AI really is.

[@andyt1313]

I’m not a programmer but am technically savvy and am obsessed with all things surrounding this AI revolution. Big thanks to people like you who bring me back to the ground. 😊

[@cristighn]

Now that is a great technical explanation Carl and appreciate your videos in general. There are a lot of startups that want to get attention and more money from investors. But that does not mean it can create a great product. Automation of code, low code is already in the market for decades. If you need to create complex apps you still need to learn a lot, have experience and actually do it yourself.

[@josh-rz3uq]

I'm so glad I found your channel. It's so refreshing to see someone rational and levelheaded on a site full of tech bros who uncritically swallow every last drop of Silicon Valley hype.

[@haraldbackfisch1981]

Just recently watched a video on how much water AI is consuming... great to have another wonderful video that shows how careful we need to be with this iteration of AI... if nothing else, it is insanely wasteful.

[@smallclips4164]

Congratulations guys!! We are safe for next 10-15 years at least

[@Cha_HCM-je9qe]

It will be a huge disaster for a company to get rid of true SE and just use AI. I 100% agree, AI is just a hype!

[@kas90500]

It is huge disaster now, let's see how it is in 2 to 3 years. If it is this bad in that timeline, I will come to agree with you. See you in 2027.

[@Cha_HCM-je9qe]

@@kas90500 Bravo a true hype-believer!

[@kas90500]

@@Cha_HCM-je9qe See you in 2027

[@nicosoftnt]

@@kas90500 You will probably say: Well give it another 2 to 3 years. And so on, and so on.

[@InnocentiusLacrimosa]

​@@Cha_HCM-je9qebravo a short sighted individual. AI is not "just hype". The person who made this video never claimed that either and actually states differently saying "I use a lot of AI tools" and names Copilot, Stable Diffusion, ChatGPT etc. Thus you stating "I agree 100%, AI is just hype!" is falsely putting words in his mouth. You are arguing either in bad faith or incompetently.

[@gifmesome]

This video points out something very prevalent in modern culture overall - the lie. All the podcast ads, the many crappy ai tools, entertainment shows, journalism, art, the lie is just taking over everything.