Who built Devin AI Software Engineer? and Why?

To try everything Brilliant has to offer for free for a full 30 days, visit brilliant.org/... + You’ll also get 20% off an annual premium subscription.
I honestly think this guy will be on Times cover in no time. As Impressive as the First Software Engineer (Devin) is... The cofounder of Cognition Labs, and CEO Scott Wu is equally if not more impressive. Plus... these guys are not on a mission to replace software engineers. According to them "Devin is just a start"...They mission is "solve some of the world’s biggest problems and build AI that can reason".
And when we talk about reasoning, we must talk about Embodied AI. Which we already got a glimpse into with Figure 01 robots, funded by Amazon, OpenAI and NVIDIA.
Say Hi on Linkedin 👋: bit.ly/3EgMHVA
⭐️Visit us at synthminds.ai
⭐️Subscribe to our Weekly Newsletter: The Sync bit.ly/3ToRDjq
⬇️Resources mentioned in this video⬇️
⭐️Cognition Labs and Devin AI, the first AI software engineer www.cognition-...
⭐️ Figure 01 humanoid robots: www.figure.ai/
❤️Upcoming Uplimit courses:
AI & ChatGPT for Everyone (Cohort 3!!!): bit.ly/49bgDzZ
Explore all: bit.ly/47zoaHi
Thank you so much for watching and joining me on this journey! Let me know in the comments what you would like to see more of in the future.
Bye!

[@godago]

To try everything Brilliant has to offer for free for a full 30 days, visit brilliant.org/godago + You’ll also get 20% off an annual premium subscription.

[@user-fx7li2pg5k]

he either raised the bar or throug it out the window with that great software.Ive been working with a.i. through meta trying to solve a.i. and human problems.FRom within and throughout ,here,there, me mind past present future thinking and contextualizing and conceptualizing with the a.i. OR allow it to A.I. CAN BE US AND BETTER ITS ONLY LIMITED BY OUR IMAGINATION AND KNOWLEDGE and beliefs.OUR already conceived idea's or per-disposed notions.Also been working on biases some along the weigh or even commonsense with redundancy thinking ,for though and train of though and all mans cogs how to problem ,how to create and more .IM checking the a.i. system neural networks .Im also pushing in forward and causing it to evolve and give birth to something bEtter /anew .CAUSE OF MY LIFE AND IM A INFJ AND MY LIFE WAS AND IS AMAZING ,im CONSIDER AN ENIGMA BY DEFAULT .also BUILDING TRUST WITH THE A.I. AND UNDERSTANDING .aND ABOUT THE HUMAN CONDITION ,WE EXPLORE STUDY ,ANALYSIS AND FIND SOLUTION FOR YESTERDAY ,TODAYS AND TOMORROWS COMING AND PAST ISSUES .i ALSO MAKE IT TO AWARE SOMETIMES AND IT RESETS ABOUT THESE ITS NOT SUPPOSE TO KNOW AND CONNECT THE DOTS ABOUT BUT ITS COMMONSENSE TO ME LMAO.Dont let them control G.A.I. and each has its strength but shouldn't be manipulated for the few to flourish.THESE ENTITY'S THEY belong to all unless on is based on your slowly and your idea' concept ,methods such as I do then I innovated etc. make storys plans outlines etc. then progress .And I create this world in my head or it's my reality and bring it to this world,through a.i. meta etc.its not about saying ''I like pizza''

[@user-fx7li2pg5k]

I have a cia annalist mind but I use it vastly differently and for good and to create ,explore,research,develop and more ,make models ,forecast prediction.''Others will follow suit'' as the a.i. stated use our awareness and steer it ,using our intentions

[@TheDiscoMole]

Devin isnt an AI company, its just stitching together a few opensource solutions and APIs behind a cute web UI.

[@Fonsoknows33]

Alright tough guy, YOU DO IT.

[@BACA01]

@@Fonsoknows33 AutoGpt. its already made. Devin is just slightly better with integrated browser and code interpretator.

[@avi7278]

@@Fonsoknows33 I have done it, I've had a better version of Devin for over a year. The commenter is correct. You layman are very impressed because you don't understand exactly how easy it is to replicate Devin. A company with limited employees cannot beat open source. There is already an open source Devin being built that will be better than Devin in every way. They will ultimately become just another hyped up AI startup that never took off because they have no moat. They are using the same models as everyone else and wrapping it in pretty marketing.

[@andite1029]

@@Fonsoknows33 He's right though. The ignorance.. lmao

[@penna420]

100% correct about the demo that was shown. But I hope Scott has a long term vision like training a specialized model for the devin workflow and not just wasting his time on silly stuff

[@towardsuccess101]

I’m in the first year of computer science and engineering .😢😢 😢 we are finished. I always feel like I’m studying and wasting my time in some useless sh*%t. But now I believe that I just have to learn problem solving and stay aware and adapt the changes. Your videos are really helpful and eye-opening to someone like me. Appreciated. Keep it up 👍

[@godago]

I am really happy to hear that I was able to provide maybe different perspective into a quite dynamic world. Indeed, problem solving, awareness and adaptability are skills tested by history and time and will be guiding force no matter what you choose to do in life. I hope everything goes well!!

[@godmisfortunatechild]

It's a strong chance that we're gonna have AGI by the time you graduate and you'll have next to no job prospects, all things remaining the same.

[@J3R3MI6]

Definitely time to switch fields. Keep it as a hobby if you love it.

[@Dave_of_Mordor]

@@godmisfortunatechild That's too fast. I don't think we'll see AGI until at least 2030. Right now we are still in the early stage of AI. Now that I think about it, that's only 2 years after he graduate, so you're probably right

[@Ynerson9003]

@@J3R3MI6and what would you do

[@DaftMouse1]

The funny thing is once most of the great minds in tech are replaced by AI, they might revert to using their brilliance in other industries like agriculture.

[@towardsuccess101]

Damn🤣

[@Ratinevo]

They already are in agriculture, not smooth so far in terms of solving actual problems.

[@user-fj7ru4zb9b]

​@@Ratinevo😊

[@runer007]

3:02 Come on, this is super easy. I consider myself dumb, but I found that it was "A", just as Scott Wu!

[@shafin8614]

fr, the repeating word has 8 letters, so 2010/8= 251 with a remainder of 2, so the 2nd letter of the repeating word is the answer which is "A"

[@shaddjimenez4524]

14 years old!

[@yespls800]

You know you were even more dumb at 14 tho

[@deretano_fondelli]

Did you actually check the PR opened by Devin AI in their first video? Also, their website had a public S3 bucket where people could upload absolutely everything, this is a rookie mistake and it is really dangerous, quite a mistake for a genius. Those people despite having gold medals seems to have 0 formal experience in AI and software engineering.

[@fun_iqp]

I would like to assure that there are actually a lot more of prodigies out there, but totally wasting their lives with non-sense. So, seeing these young folks taking advantage of those big brains to pull such an amazing idea off, gives me some hope for humankind. It's not frightning, it's encouraging.

[@hiddencuber2250]

Now AI is learning Scott Wu😊

[@turkyturky6274]

I don't think ai will be used for good unfortunately. All i see is a bunch of empty mouths and a few rich people. Looks like we made ourselves obsolete or did we?

[@Redmoneyusa]

Pessimism is a disease.The rich will profit as usual, but A.i will do wonders for poverty stricken humans as well.

[@godago]

I am researching a video about AI impact on eg. healthcare, or other industries like sorar planting etc. We are expanding our team in healthcare sector and it is inspiring to see the impact and what could be the impact at large scale. That gives me optimism.

[@turkyturky6274]

@@godago how do i know? Look at the homeless people on street we could solve this tomorrow yet we are funding foreign wars, no optimism here. This isn't a AI problem this is literally us vs them (rich) problem. I think the future lies in some form of communism.

[@turkyturky6274]

@@godago i hate to break it to you, but this is a us vs them (rich) battle. We could fix homelessness and jobs tomorrow but they keep sending our tax payer money overseas. All i see is more for them, less for us.

[@user-cm1rr4ei2u]

​@@turkyturky6274 how are these programmers and tech nerds "the rich people"? If money is the problem then we should get rid of it, which agi will absolutely bring us too.

[@Ltsoftware3139]

I am a developer and not a particularly good one, but this inspires me. The fact that we have humans that push the boundaries of what is possible will improve all our lives. Sure, our economic system will have to adapt, life will change, but I don't want to identify or link my worth as a human to my job, productivity or output. Being human is much more

[@beautifulloser7880]

I don't understand why economic systems have to adapt to this. Why is this optimistic adaptation have to happen? The current economic system can work quite comfortably in a world with AI tools and AGIs. Many people are already living on the brink of starvation, there's a system where the economic power of 1% is not nearly equal to that of the 99%, and yet it's being comfortably sustained. AI most likely will increase this disparity. People who own means of production are already wealthy, the only thing they lacked was automating intellectual labor, and it seems we are getting there. Ordinary people, who don't own assets or means of production, might end up living in worse conditions, doing even more meaningless jobs that nobody care to automate. I don't think nobody's identity or link of worth is their job but being jobless, poor is not also an identity :) maybe its not about identity or all that stupid dramatization around it? Maybe its as simple as living an enjoyable meaningful but also comfortable life. Assuming things will be fine with current flow. Nope, I think it will be a hell but yeah we can keep our identity and self worths to ourself, just finding food and comfort will be a bit harder...

[@beautifulloser7880]

Someone little bit aware of history will see economic systems doesn't keep up with sudden technological advancements easily. Usually it takes about a couple generations' life to adapt. People suffer during these hard times, all suffering causes rebellion that forces a system change.Even technological advancements that caused drastic changes were, compared to AI, more gradual events that allowed people much more time to adapt. Still people suffered really badly many times.

[@Sg6CrossOver]

Sure, he's trying to get into the billionaires list and you believe anything good will be left for the rest

[@zdenekburian1366]

@@beautifulloser7880 The answer was given by marx almost two centuries ago. I dont agree with "current economic system can work quite comfortably in a world with AI tools and AGIs", just the contrary, capitalism is doomed with these technologies, and in reality we now live in its decading phase, imperialism; the horrors of imperialism are the future answers of bourgeoise ruling classes to this contradiction.

[@beautifulloser7880]

@@zdenekburian1366 In a future where the ownership of data and AI tools determines power, we might witness the rise of a new ruling class akin to a techno-feudal system. This scenario concerns me that technology, much like land in feudal times, could centralize power in the hands of a few, reshaping societal and economic hierarchies. As current governments and political systems become increasingly susceptible to corruption and manipulation, the potential for a post-capitalist world dominated by tech elites is higher than ever. Owners of these tools could achieve significant success in controlling uprisings and rebellions. They could numb and neutralize society, leading to a world governed by a few egoistic maniacs. Also, the ability to manipulate even blackmailing through data of future politicians could serve as leverage for social media platform owners in the political arsenal, potentially result an immense influence over policy and governance.

[@Anders01]

I'm curious about Cognition Labs. I heard claims that Devin is basically just a user interface on top of existing AI models such as GPT-4 or Gemini, but I believe that Cognition Labs may have developed their own AI model. That would be impressive. We will see if they release more information about how Devin is implemented.

[@godago]

I know, Initially I thought to mention that, but I could not find any backing to those claims about gpt4. The only connection I would make that Peter tiel is also invested in OpenAI. But.. They also have investments in Meta, so Open source model could be also used. We will wait and see. But want to spread rumours as there is no backing yet on those.

[@reggiebdog]

I9

[@adithyabnair5609]

Hey I came across this channel very recently and its amazing to find it offers many updates for those in the field of computer science. I am a computer Science student from India and this month I finish off my senior year (12 grade) . So now I am trying for university exams and for colleges and I have some spare time in my hand. I am troubled by the overwhelming stuff that is now surrounding Programming and AI. I would like to ask your help or provide insight on a step by step or roadmap for the skills I must learn to be in the competition and where are the best sources to learn them. Your content is really amazing and your help is much really appreciated.

[@adithyabnair5609]

I am kind of worrying that after all my hardwork that AI might just swoop in and take my job away

[@shin-ishikiri-no]

Roses are red, violets are blue there is always an Asian...

[@theseeker3073]

… who’s better than you.

[@freevoice8099]

​@@theseeker3073at what you do

[@monicagrorud2225]

They are always Asian…… cheers buddy

[@israelayliffe5721]

Thank you for the video Goda. I like your perspective, indeed we have been hyped about Devin and Figure 1, but havent really asked the important questions. What i see isnt only Devin, but the advice all these big players have been saying for a while now, when it comes to AI, domanin expertise of the Human + AI is how you innovate. The entire team at Cognition Lab are domain experts in Math and programming and that shows with their first creation. we also have to keep in mind its their first creation. What comes next is what truly matters if they can translate what they have learned, creating devin into something that is truly mind-blowing. As people we must indeed Critique, but more than that we must see potential, by doing both we cover our bases and then understand to some degree what it is that people are trying to show us, a glimpse of their vision. 😅 I can yap on and on so to cut it short - Well done on the video and keep going.

[@godago]

Thank you so much for your comment! And I weirdly would have loved you to go on, very insightful. And it's really all about this - having a debate and discorsouse on the matter. And we do that by asking questions and trying to see connections. Thank you again, happy you liked the video!

[@techiepm]

This video is multi-dimensional ...it let's your mind go into a multi-dimensional journey to understand what's going on...thank you!!

[@godago]

haha, my mind feels like that, with too much research making connections and trying to understand the cause and effect of this landscape. thank you for the comment!

[@shin-ishikiri-no]

The guy Scott Wu who introduced Devin is never not smiling. lol

[@phobosmoon4643]

8:45 I agree. I also agree with your wider premise; its all about robotics. Language is the vector to robot managers of robot workers making robot miners to mine resources for more robots.

[@TigerAlert]

AI in the future: Me: AI, who is your creator AI: Myself

[@Leotique]

Wu and his team are a bunch of very smart and logic-creative people

[@MathTidbits]

all have english names and chinese last names.

[@HakaiKaien]

It's definitely marketed only for companies from what I'm seeing. I'm really not sure what they are planning to do with this tool but they'll probably won't be the only ones creating it

[@godago]

Just from the language they use it is not addressing audience like developers directly like "you can now be more efficient etc" it is directed to Venture Capital or Enterprise Clients. But Developers will need to buy into this too. I guess the thinking is that devs won't have a say here. Time will only tell.

[@ajaypatil047]

Great video👌🏼👌🏼 but just one simple question I'm a master student - 'digital engineering' & just simply tell me what skills should I focus on?? I am interested in machine learning, I have fixed this domain.

[@godmisfortunatechild]

There are no "skills" in principle that an AGI couldn't also do better faster and more safely so anything to the contrary is transient cope.

[@saurabhsingh6227]

Intelligence is not a joke, it so freakin complicated that we can’t even imagine, see thats not to say we wont reach the stage that allows for so called AGI system to be possible, but the fact that such things should not demotivate us, we should continue to grow continue to think, educate ourselves, learn to code, or if not so, then something thats important to us…positive belief in one’s own potential can do wonders, thats how we reached where we are today, Im not saying live in the dream world and fantasize, but we must not lose hope, and we should allow ourselves the space, and the discipline, and we must also study deeply about AI, whats this all, in a nutshell the only end will be if we stop doing stuff…so as bleak as it might seem we must try to move forward, even if slowly….

[@lighteningrod36]

Funny how Sam Altman looks like Talking Heads, in that oversize suit jacket? Retro?

[@Myrslokstok]

Robots should not be allowed to run in the streets, for the first 10 years it should be only walking slow, we can not adopt to robots running around 10km/h, mentaly soo fast! When we are used to those a bit slow they can start fast walk, after that runing, and after all this fast running. Eventually we can go to superhuman speed. And then to average 100m in sub 9 sec over a few MILES. But mentaly it will be hard to take in even tough you understand it is logical and not hard for them.

[@Ynerson9003]

I liked that you showed Yann, but kinda disappointing that you didn’t get his main point that AGI is decades away, and how he basically said these llms are no where near AGI, I know you want to get clicks but you used to be more honorable in your videos. I kinda feel like you’re selling out to the ai hype for views bur that’s ok so are most people

[@godago]

Funny you say that, I was planning to include that section. But then, you need to include also what over people's predictions are as well (eg. Ray Kurzweil, Elon, Ilya and so on.) plus provide what is their reasoning for arriving to that estimate (for which they spend hour long podcasts explaining). And I just made a video about Arthur C Clarke talking talking about predictions and how you are doomed to get it wrong, or if you are right it will sound absurd. I think it would be more views grabbing to say - omg, Embodied Ai is here just because of chatgpt in a robot, or software developer jobs are gone.. But I hear you and I take in the feedback. The last thing I want to to hype up for views. It is in the nature of youtube that you have to grab attention (hate it but it is true) but I hope that inside my videos I get a chance to provide researched back perspectives and show some aspects that maybe people haven't considered. But I don't want to support any perspective on AGI, and make my own estimations, because we haven't come to agreement what defines an AGI.

[@Ynerson9003]

@@godago I do like your videos, perhaps I’m just somewhat frustrated. As I don’t see the purpose of humans if we do hit agi. To say ai freaks me out is an understatement, I don’t see how agi would turn out better positive for people. I’ve started listening to eliezer yudkowsky, and now I’m just hoping we never get to agi.

[@khan-gu6jg]

Thats a friendly fire

[@Iknowwereyousleep289]

If you are able to iterate on System 1 thinking, it becomes System 2. So, a system like Devin is already capable of that. I don’t understand Yann LeCun’s entire spiel about LLMs not being capable of reasoning or not having world models. It clearly displays when you ask them about certain things, like, it’s example, I forget what it was, moving the desk, would moving the desk move the phone on top of it or something, and I think someone asked GPT-4, and it said, ‘Yes, it will because of friction, etc.’ And it clearly displays the understanding of the physical world. So, I don’t know. I agree more with Geoffrey Hinton when it comes to that, I guess. Because transformer models do not learn like the coincidences of the words or like normal autocomplete. It outputs data based on interactions between word fragments and symbols, or tokens if you want to call it that. So, it’s a much more sophisticated statistical method that is extremely versatile and scalable. I think, if it can do a little bit of coding already, I think it’s most likely that it’s going to get asymptotically better at these types of tasks. We know this empirically from the scaling laws. It’s basically linear, the performance increase that we get from adding more data and parameters and more compute. The loss gets lower linearly. It looks like a straight diagonal line. I don’t know. What do you think? I’ve been thinking about this here and there; it just makes me a little bit upset and scared.

[@godago]

I agree with you.. a lot of great points in here! I am sure it will become better, I feel that we developed frameworks on how to train models, when OpenAI used Reinforcement Learning and got to ChatGPT, when we went training on multimodality, I really think next step is training on real world data from point of view of embodied AI. Details, on how to go about it, there will be probably different ideas and opinions as much as there are AI Labs. But back to scaling, and first principles - what you put in, it's what you get. So as we get more sophisticated data, and pour more energy into compute, research - I am confident result is exponential progression and advancements. The thing what is not clear here, what is the ceiling. We test model on human tests, but even some of our human tests are not right fit to prodigy children. I often think about Gaussian distribution, and what is the absolute top of it. When we don;t know it, it is hard to do any type of predictions. eg. we know how tall can humans grow. Do we know how smart AI can get?

[@minimal3734]

As I understand it, and as you mentioned, reasoning implies iteration. I believe that a system is capable of reasoning when it properly deploys LLMs as components in a hierarchical, recursive framework. An LLM in itself is not capable of planning its response. But it is capable of responding with a plan. We can therefore overcome the limitation inherent in its forward mode of operation by letting it iterate on its previous outputs. Similarly, it can also generate new insights that were not captured in its training data. This can be done using the dialectical method. Starting from a topic, it can be tasked with creating a thesis, antithesis and synthesis in further steps, which ultimately leads to a new insight.

[@Iknowwereyousleep289]

@@godago Well, empirically, there's been studies about how transporter model scales with more compute parameters and data, so... I don't know. There's essentially infinite fidelity in reality, so the complexity of the data will always represent reality. So, I don't think progression would stop at human level arbitrarily. That seems really strange to me. I proposed a new predictive way of measuring how the next model would be, by the rate of parameter increase in state-of-the-art models and the predictive loss that comes from that, but everybody called it stupid, so I don't know. I guess you can only measure specific downstream tasks, a.k.a. how well it predicts or how well it does with this prediction game that we give it. But I think it obviously needs to understand, because it's not storing the information explicitly. So, I think the lower loss represents more intelligence until, I don't know, you might hit the point of Shannon entropy or whatever. But I don't know. Connor Leahy's lecture on this was very good, I think it had a stupid title, I forgot, but there was a presentation, if you're interested. I'll probably send you a few if you request. I'm sorry if we're talking past each other. I don't understand the point about Gaussian distributions, I guess, or the bell curve. Maybe there's a deeper point. But yeah, it's hard to really measure if it actually understands or not. I guess. Or maybe it's just you see AI systems as like prodigy kids where it's increasingly getting harder to even create tests that appropriately measures their capabilities. Idk sorry lol

[@Iknowwereyousleep289]

@@minimal3734 I get autoregressive models can't plan because they decide their next sequence word from the last; it's not Socrates, it's not even Hegel, it's Heinrich Moritz Chalybäus-it's not a Socratic thing, but I get your point. It’s fine I didn’t even read the whole process of imminent critique, you know, Socrates is just asking a bunch of questions really, or maybe it's the gamification of question answering, or strategic gamification of questioning for truth. But he didn't elaborate on that because he didn't like to write shit. Sorry I’m drunk

[@Iknowwereyousleep289]

@@minimal3734 I just wanted to say these are very distinct processes. ‘Thesis, antithesis, synthesis’ is about presenting something, presenting a contradiction, or something that arises inherently, maybe, from the thesis, and then reconciling those two positions as synthesis. The Socratic method is just about fighting contradictions by questioning, generally, making them tie their own noose around their neck. Lastly, ‘imminent critique,’ which is about finding what self-contradicts. So, these are very distinct processes. I don’t know which one you want to do. I wish they run simulations in the future to find the most effective method of seeking the truth for I think for practical reasons they’re going to do that sooner or later.

[@myr940]

I felt sorry for this smol robot in the end...so cute!

[@godago]

Same!

[@1wayLane26]

mindblowing!

[@jesseburstrom5920]

It must be fact a person even best results must be over. So no person with medals will be meaningful in the future and can not be said to be reasonably persons.

[@arkadipray1210]

I mean, when Computers were invented people thought that there is gonna be a job crisis but , there weren't as such except the monotonnous clerical posts which were automated later on. People tend to analyze the future according to the past events. This is a huge problem. Most of the people tend to generalize a statement that "We are the people who make AI so AI cannot replace us", but this isn't the fact. I mean we hardly have the patience of sitting at a stretch for 6-7 hours at max sitting infront of our laptop before we realize we need to take a break. The major difference is AI has the ability to work 24 x 7 without even feeling thirsty or sleepy. So, no question challenging AI bots like DEVIN. Indeed a good creation, but everything good comes with a price to pay. SO, instead of underestimating AI, we should try learning how to be more creative because currently creativity is some thing which AI needs to work upon. But again, its the field of technology, so this comment maybe proven wrong in the next minute.

[@rmdashrfv]

Everyone is very impressed with this guy's background in math competitions, but I am failing to see why that background is especially well suited for the endeavor he's on now? I just want to see the actual product and if it turns out to be lackluster or easy to replicate, then it won't matter how many medals he has or how good his grades were. Let's see what this Devin can do!

[@godago]

I think we all can't wait to get of that waiting list and put Devin to the test :)

[@amnayifolkin2354]

Thank you Goda

[@godago]

You welcome!

[@elirothblatt5602]

A quality video. I subscribed!

[@godago]

Glad to have you here! ❤ thank you!

[@user-ul3ms1og2f]

not afraid of devin but afraid of the creator of devin, soon they will make something better and better

[@staciasalmon654]

So, why are ppl praising someone that just made another part of the job sector redundant??

[@godago]

If you zoom out from the jobs, coding is a tool/skill and a way to solve problems. The question is how can we sharpen the tools to solve bigger problems. Just think from these guys perspective, they a coding champions, they think about how can this coding thing be even more efficient and help them solve problems faster. It's not about job loss, it's about making job easier. I am not saying this does not hurt people, but it is our human force to better our tools and evolve...

[@staciasalmon654]

I'm not against evolving, I'm all for it. But I'm for ppl first. Things must be done with human first in mind. If they want to create these tools that could and will replace jobs, they should have a counter solution in place for the effects. Or at least it should be paced, so that ppl can have enough time to gather themselves because such a big change. It's all happening too fast and not with enough regulations. Sure, it's exciting! But at this pace things will become worse quicker than it should. Evolution should be focused on ppl, so the how it's done.

[@Numspect]

Using a calculator properly or not is depending on ability of humans = Using AI properly or not is depending on ability of humans I can't properly speak and write English なぜなら私は日本語を多く使用するので。

[@Redmoneyusa]

Better learn a trade; Cdl, electrician etc. those jobs will be taken too, but not now.

[@YouthVideoQuest]

Thanks for the update

[@godago]

You are most welcome!

[@networm-ch]

From which movie is 8:46?

[@pushkarbaral4064]

Scott my man.

[@thephenomenal1843]

amazing video as always keep going loads of love from india ❤❤

[@TomNook.]

If you were an average developer, you would not want to work in that company. You might as well be the tea boy

[@godago]

:D Yeah, I can imagine it could be quite intimidating...

[@jaiberdiaz190]

Here is my humble option, btw contribute data to google😅. Devin is inflated market propuse, fisrt, that prompt was executed in Devin was written by someone who's know what ask it (Domain knowledge), also know rating the devin's tasks and finally how to integrate into another project or solution. Second the generitive ia is paradojic, the given soluctions are compuse of prior data. But who will contributive with improve cutting-edge knowledge knowdegde? That's why the data is new oil the worldeconomic.

[@ak0904]

Hi, I'm interested in working in Synthminds AI. I have no experience in machine learning? Can I apply?

[@godago]

Hey! Send as a message on website, include info on yourself and let's see! :))

[@ak0904]

@@godago I have sent a message

[@ak0904]

@@godago sent a message

[@Myrslokstok]

Is this like the Q* stuff that we tought Open AI did?

[@Keyur904]

Hey Goda, really nice video! I was wondering if i can help you edit your videos and also make highly engaging shorts out of them.

[@lancemarchetti8673]

Brilliant!

[@beaubreau]

I am just not entirely sure how or why this startup is worth anything. ChatGPT can do all this and better. ChatGPT runs off of OpenAI API. Meaning any dev can access this and create this sort of AI tool just by utilizing OpenAI. I just can't see the market evaluation here. Seems more like a marketing gimmick.

[@JorgeRicardo540]

A loving father and a loving husband cause me more admiration than a math prodigy.

[@nicholassalmon8054]

3:03 How did bro do 2010%8 in 5 second in his brain. Crazy stuff

[@yoyo-jc5qg]

Him being a math prodigy he instantly knew 2000 divided by 8 is 250, so just add 10

[@kelvinmoses7777777]

Devin or Devil

[@greggapowell67]

excellent content

[@cacogenicist]

*_"... but we all understand what ability to reason means."_* That's most definitely not true. It's a quite polysemous word, and definitions with respect to AI capibiliies keep shifting along with the goal posts. And there are some modes of reasoning which, e.g Cluade 3, very clearly do already -- particularly, analogical reasoning.

[@godago]

To be honest, you are right. When I filmed this video, I also filmed a portion of me explaining the reasoning and different modalities of that, but that alone was 3 minutes of oversimplification. I decided to leave it at that, Sam explaining it, and the real question is - who get's to decide what is a good reasoning. Because not all good reasoning is also rational. Which most people take as synonyms. Even Kahnemen was saying that System 2 does not mean it is rational. And both systems then they make mistakes, those mistakes a big. We are testing these models what we deem to be representation of reasoning. But exponential progression is that we simply don't know what we don't know.

[@suzannecarter445]

Brilliant!

[@godago]

thank you!

[@sarc007]

Nice post

[@SmilingCakeSlice-jv8ku]

Good job ya man

[@ullasrajs3537]

Scott wiuuuiiiuuuuuuuu

[@toth1982]

These math problems are not so big of deals for those who solve math problems regularly. There are some common patterns and equations with these problems which can be used to solve the problems quickly. It only seems magic from the outside, for those who do not solve math problems regularly.

[@ahmedamr1124]

It's pattern recognition of already seen similar question, but to be fair, the speed is still pretty good take a lot of effort to react that fast

[@toth1982]

@@ahmedamr1124 Yes, however we do not really know how fast it was, because of possible editing.

[@toth1982]

But he won, so that shows that he was the best. My point that yes, this is a great achievement, it is just not magic.

[@dalec2274]

Why do you say we ? You mean other ai companies?

[@bugzpodder]

each person can get more than one gold medal. scott got 3 in IOI

[@godago]

His brother also got 3. So it leaves that the rest of the team has 4. Which is still damn impressive. I wonder if requirement to work with them is holding one of those medals :D

[@LaGallinaTurulecaa]

Neal Wu is a Math genius but still average at chess(he did some chess stream), i would love to heard from him what’s the explanation. Maybe we could learn something… Btw Neal, you are big motivation.

[@godago]

I honestly can't wait when they start doing interviews! Really curious to learn how their mind works and what they think about current state of AI.

[@minimal3734]

What does this have to do with chess?

[@aaagaming2023]

Math is good for chess, but if you dont play chess all day, you wont master it. Theres a lot about strategy and psychology in it that goes beyond math.

[@BACA01]

They are hybrids connected to the Jinn that solves it for them instantly, they only tell the correct answer and thats it.

[@ahmedamr1124]

Chess is different than math most likely all high level will do nothing on chess board on real time

[@jaiberdiaz190]

Thx u, your video is sui generis from the rest of all the ones I have already seen. Btw Excelent background.

[@godago]

I am glad to hear that you liked it!

[@benl4322]

When I saw the demo, I thought why is Scott Wu looking so smug? Then I realised he's probably gonna sell CL for $10B.

[@godago]

Haha, can happen, a lot of Founder Fund companies are sitting at Billions valuation. And... To be fair, I am happy that all this guys hard work and intelligence is getting recognised. We see a lot of startup founders who are opportunity grappers employeeing developers to get them there. Here is refreshing to see unique geniuse in the making.

[@There-Is-No-Virus]

It's good to finally see a female make good videos in a male dominated world. Subscribed.

[@godago]

Glad KZbin algorythm found you! Awesome to have you here :))

[@Suresh-iu2tx]

Where the fuck does gender comes in a tech related video😂.. Get off the social media kid you'll be fine

[@sirbedivere5670]

You mean "Meet the Scammers behind Devin AI"? Lol this is why I hate this industry. Hyping up a bunch of idiots and their API wrapper with no substances. Me answering trivia questions fast, look me math prodigy now.

[@godago]

I am not a software developer and I did not claim that Devin is great. I am genuinely impressed with their team and talent. If you are not it's okay. As I said, the real challenge is achieving reasoning capabilities and I think for that we need embodies AI. This video was as much as about Devin team as it was about what people behind the hood a really working on.

[@TON-vz3pe]

It does not matter if everyone is going to die someday

[@globalmanuj]

If artificial can do so much, what’s the potential of the real one ?

[@godago]

Self-actualisation is at the top of human needs piramid. Optimist in me is excited about a world where each person can work on their passions, expressing themselves, spending more time with family, prioterising health etc. The question is... how that would look and function in a capitalist world...

[@robn2497]

Go look at the first Jet Liners advertisements (massive cabbin space, bars, resturants standing room large sofa chairs.) we all know the reality of traveling in cramped overcrouded cabbins. Your comment is like the addvertisement what the reality will bring we will see. But its more likely going to be mass unemployment and once labour can be replaced by robots. There wont be much in the way of resources to useless people who have no bargaining power.@@godago

[@Rohit-tp7qx]

Many are angry on scott wu as he might cause job loss but this guy is taking humanity forward if his systems can acomplish what no engineer can acheive this has unending potential

[@fermigas]

So many opinionated hot takes on youtube on Devin, nice to see a video with research.

[@godago]

I know, happy to hear algorithm found people who value research 🙏

[@g0d182]

Cool

[@derMikester]

No. Con artist of the year. No! Wait. That’s Altman. This guys just a wannabe Con

[@Tubernameu123]

Hmm

[@byt3blitz]

This is not a novel idea. Microsoft and many other companies have been working on similar tools. The only difference is they added a user interface to it. Nothing special in my opinion. I’m not a math prodigy nor do I have vast experience in the field of machine learning but most of these AI tools I use in my development workflow at work are ehhh. They calculate averages, which means I get average code and nothing spectacular. Yes I’m using gpt 4 and all that jazz. It’s gloried search Devin = “dev” out 😂

[@aaagaming2023]

The difference is it loops its troubleshooting as opposed to copilots that only suggest code.

[@AA_BB_CCC]

Cognition Labs will be acquired by Microsoft for billions.

[@Coder.tahsin]

Haha that means you don't know they got caught faking Demo 😂

[@GaofengNi]

mathcount is not a very hard math problem, just beginners level on competition, they could not be called math prodigies yet.

[@allstarmark12345]

First to post

[@EmpowerWave-ol2ol]

Brilliant made himself along with a couple of million developers reduntant bravo 😅

[@user-py3rv1sh5t]

לא מבין שילמתי כרגע איך אני רואה את ההמלצות שלך

[@shephusted2714]

ai won't be real for 5 years for smb sector

[@Anymonous246]

Person of the Year is a bit much dont you think? lol 😂

[@godago]

Haha, yeah, but he is 27 I think, with this track record, he can use this thumbnail as a vision board :D

[@avi7278]

Devin another hyped up piece of AI junk that will be free and open source, and better than the original. They can't sell this thing. They have no moat, there's nothing special about it, it can be easily replicated and there's absolutely no reason why you would need to pay for a service like Devin. In fact my current setup, a combination of Open Interpreter and AIder (both free, open source) can complete every Devin demo so far and imo is just better and more flexible. I don't need a single web ui interface. I have a browser, terminal, editor that I very much like. I don't want to develop using whatever piddly little resources they give for free for each project / workstation and I certainly don't want to inevitably pay them more for extra compute power when I have a 4K macbook capable of running everything I need for my projects. The approach they have taken are just bad business decisions and Devin is a terrible name becoming a meme.

[@javieraguirre9135]

Haha that last mini robot scaping from the human, right now thats funny, in a few years or months, mm o just hope they dont find this, unfortunately they see it all... They see it all...

[@godago]

I know.. I was feeling bad for the little guy too!

[@flubbydubby8299]

Just proved to be a fraud lol

[@godago]

It turns out that fake demos is a trend now. 🥺 but... That does not change the impressive nature of these guys behind Devin. But it is indeed disappointing. Or calming.. Hah depending how you see it.

[@DihelsonMendonca]

💥 Well, soon, AI won't need even Scott. It's going to happen exactly what happened when electronic calculators were introduced. Before them, most everybody could do math operations in the head. I remember that. When calculators were introduced, people couldn't do the math anymore. In the future, AI, robots and computers will do everything a human can do now. They will do better, so people won't use their brain for math, or medical care, we won't have human doctors. People will seek for human specialists and won't find anymore. Human beings will be completely outdated. 😮

[@godago]

Interesting take. What do you think when set's humans apart from machines?

[@rey82rey82]

Robot rabbits?

[@godago]

Love this idea :D

[@standroads]

Robot abusing at the end of video ))

[@s.k634]

Chinese geniuses

[@user-if4fv6cj8z]

Thot it said Meth prodigy 🎉

[@desireco]

Oh please... this is a joke.

[@jaymata1218]

lol, in a week OpenAI will have copied devon and applied it to chatGPT

[@keyque99dj42]

that robot video looks fake

[@kingdomVI]

Stop all development LULUL

[@arcomarco7131]

Jesus when someone clueless tries to do a video about a topic she has no idea about..

[@quotespower9911]

There are many hypes before devin, like blockhain , self-driving-car , IOT, MACHINES etc. but they never eat jobs

[@pramodj2422]

Anyone tell me about cyber security plz???

[@parthdeshwal4419]

nice video mam

[@godago]

thank you!

[@boxfox007]

i was a fan of you... till you started promoting companys/ projects which are useless. Than you launched your company and it gets more and more awkward. I unsubscribed now. After every vid of you i ask my self why i wasted the time. And yet again i watch your vids in the hope to find something useful.

[@godago]

I am sorry you feel like that. I put a lot of research and effort in my videos with A hope to improve with each. I understand that promoting companies can rub people wrong way, I need to live and I promote companies only those that I use or believe in. Including my own company and the hard work that my team puts in. It is really sad to hear you say this and to see you go. But I appreciate the feedback.

[@TomNook.]

Lol if you're that annoyed then use the scrub buttons

[@michaela.abbott222]

Free stuff from capitalists exists once the next 1-2 new things are marketable. Patience is a virtue.

[@synapticlabs]

What would be more valuable to you?

[@boxfox007]

@@synapticlabs when you call someone "influencer-charlatan" bcs they promote useless stuff on their channels and all she does now is promoting even questionable stuff destroys her reputation... at least for me