Hacker News new | past | comments | ask | show | jobs | submit login
What the hell happened to Devin AI?
11 points by walkerInWalls 19 days ago | hide | past | favorite | 34 comments
why do you think they couldn't capitalize more on the hype?



The last thing I saw on it was a video saying that calling Devin an AI software engineer that does Upwork tasks is a lie based in what was shared in an earlier video.

https://www.youtube.com/watch?v=tNmgmwEtoWE

"I broke down the Devin Upwork video frame by frame, and here I show what Devin was supposed to do, what it actually managed to do instead, and how bad a job of that it did.

On the whole that's not surprising given the current state of Generative AI, and I wouldn't be bothering to debunk it, except: 1) The company lied about what Devin could do in the video description, and 2) a lot of people uncritically parroted the lie all over the Internet, and 3) That caused a lot of non-technical people to believe that AI might replace programmers soon. "


It's not just non technical people. Even programmers around me are worried. Some think AI will replace programmers soon, some think partially, mostly everybody thinks that a lot of stuff will change soon.

I am personally not worried. Any outcome is great. Either I keep a well paid job and it becomes even more well paid, or AI is so good that there aren't any jobs at all anymore, which sounds good too.

Let's define soon - I understand it as "within a decade or so" in this context.


This is my belief as well. Either one of three things happen (in order of most likely to least likely)

1. AI replaces software engineers in the same way that "autopilot" in planes replaced pilots. Augmentation of boilerplate (particularly devop CI/CD, deployment tedium) sounds pretty appealing to me.

2. AI genuinely replaces software engineers. If that's true, there's no reason to believe it wouldn't soon disrupt other engineering domains as well. At that point we're basically closing in on AGI, which doomsday predictions aside, I would like to hope leads radical progression of technology and evolution of mankind.

3. AI is completely banned/suppressed Dune style. This only strengthens my position in the market.


AI genuinely replaces software engineers. If that's true, there's no reason to believe it wouldn't soon disrupt other engineering domains as well.

If that’s true, then…

How? Most engineering is physical and blueprint work which cannot be converted into embeddings and filled in the middle. Or if you’re assuming that something LLM/VLM/SD-like will appear for this domain, that’s a pretty wild guess. I even doubt that transformers are more suitable for that than now-classic boring ML. If you hope that the language will somehow “emerge itself” into building, schematics, industrial, etc behaviors, that’s also quite a guess.

Make sure you’re not coping, because transformers generalize over (fuzzy) painting and text, the only data we have in abundance. These are common attributes of programmers, writers and graphics artists and barely anyone else, and the latter two are already screwed. We don’t have readily accessible logs for abstract thoughts, professional conversations at work or inner monologues. Maybe in ten years China will realize it and start to collect it at scale.


This seems like a big assumption to me personally.

To be clear, I happen to be of the opinion that AI/ML/LLMs/whatever cannot truly replace software engineers without the true advent of AGI.

But to play devil's advocate. I'm an embedded SWE and work with circuit designers/have dabbled in some PCB design myself. Taking circuit design as an example:

From an abstract logic standpoint, PCB design is not that different from software development. It essentially boils down to taking some input and feeding it through a cascade of various discrete transformations until you have some desired output. The fact that PCB designs are captured through schematics is irrelevant, as a schematic is just a visual representation of a netlist. There are even DSLs that allow for the design of circuits through code alone. The mechanisms that make LLMs work show some level of adaptability to domains outside of traditional language. It is totally conceivable today to finetune an LLM on netlists (or even on images of schematics with the right encoder model) and have it be able to generate circuit designs. The training data exists, though it's not as plentiful as say code or english texts. I'm not an ML expert but I believe it's totally possible to have something where, for example, the text "555 timer-based LED-blinking circuit", could map in embedding space to a netlist that describes said circuit. There are in fact companies working on this exact thing today (Flux.ai)

I don't see why this wouldn't work for other engineering disciplines aswell. There's lots of public scad code for various 3d models. It's totally conceivable to build a model trained on scad or even raw 3d object file geometry and text embeddings to generate mechanical designs. Lot's of research is going into generative meshes which could be seen as an early form of this.

I think my point being that the hard part of engineering, discipline aside, is being able to take a fuzzy set of desires from a customer and a fuzzy set of constraints and being able to use the hard math and science to create and iteratively refine a product/system/whatever to make all stakeholders happy. LLMs can't really emulate this process today.


This is a great point, one i'd not considered.

People realise the AI capability is limited by the quality of the data used for training, so work out ways to collect better quality data and keep pushing on the LLM architecture as a path to AGI.


Let's keep physical engineering aside for a moment: If AI arrives at a level to replace software engineers then we can safely assume it can replace thousands of other knowledge jobs that are way less syntactically complex, from will writers to attorneys, doctors in general medicine, 100% of Excel/Word/PowerPoint white collars, business consultants, middle managers, senior managers, VC capitalists, investment bankers, etc., basically destroying the modern society. Not gonna happen.


I’m pretty sure that LLMs will never replace programmers. It’s way too hard to stitch together code snippets. Also, when you think about all the misunderstandings that arise from pseudo code, free form text input to an LLM has to be even more ambiguous. I don’t care whether someone calls themselves a “prompt engineer”.

But AI will replace programmers just like it superseded chess players.

We are putting the finishing touches to an AI that write code. And not in an LLM kind of way. It allows you to specify exactly what you want without needing to specify the logic. Check out Ac28R.com

We are excited about it, but it makes my heart ache when I think about the impact.

The world is speeding up. We are the cause but we will struggle to avoid getting swept up in it.


> I am personally not worried. Any outcome is great. Either I keep a well paid job and it becomes even more well paid, or AI is so good there aren't any jobs at all anymore, which sounds good too.

You certainly have a rose-tinted magical way of viewing the world.

You haven't thought through the consequences of this at all. I'll break this down a bit further for you hopefully in a way you can understand.

You work to provide value in exchange for a store of value; currency. You do this to spend that currency on food and other basic necessities which are dependencies for your survival, first, and then discretionary spending second.

Producers hire you so you can distribute labor and produce, so they can make a profit. They stop producing when that's no longer possible. Its an agreement between both parties which enable distribution of labor and exchange of goods flowing through an economy. Without either part, no exchange occurs, there may be a brief stalling period but when its stalled everyone's goose is cooked. It is a very fine balance, or supercritical state where deviations can cause exit conditions. Sound similar to an n-body problem; that's because it is, along with mathematical chaos (small changes in inputs create dramatic unpredictable changes in outputs).

When there are no jobs, you can't earn currency, and you suddenly can't feed yourself. You have nothing to do, no way to differentiate yourself either.

Worse, No one will listen to you because you no longer provide value, you are now a nameless worthless slave. The only thing you had was trading your time and education for money and this is now gone. Your children will be worse off because the cost of education will no longer provide benefit.

This will worsen as population grows.

Eventually because resources are inherently scarce, someone somewhere with power, will decide its just better to reduce the number of mouths to free up resources; and they'll decide for you that you and your entire family meet that criteria, along with many others, but not the majority who voted those people into those positions of power. It will be a genocide of the rational and intelligent, powered by big data and your tax dollars.

It won't be the old people because they represent the most power having started and accumulated power first, but they will eventually be on the chopping block as well.

Along with this, production systems will fail, either as a result of vandalism, or growing corruption all causing shortage, which causes unrest and violence. With no market, or medium of exchange (inflationary economies fail when debts exceed production), order wanes.

The rational and intelligent people will fight back weakly, trying to organize to survive by destroying the system of bondage and slavery that is imposed on them without their consent. Eventually everything fails as systemic issues cascade.

Because intelligent people have the capability of great harm, they will be eliminated first favoring others instead.

Are you worried yet? You should be, every single thing here logically follows the next with just a little educated background in economics and history.

The world you say sounds so good, is in fact a hellscape world of spiraling madness, and intolerable suffering with no one capable of the first step (recognition) that can stop the death march forward. Anyone that could will have been killed long before recognition could happen in ways that don't attract attention and it will only show up in the actuarial tables under mortality.

People will stop having children when they can no longer afford to.

Old will crowd out the young as medical technology improves, and then there will be a great dying and collapse.

Those that remain will have not been prepared, having developed in a disadvantaged environment and so these collapses may end in the annihilation of the civilization as a whole.

By the time you recognize the problem (if at all) its too late to do anything about it, and your heads on the chopping block along with your family and friends because you had the intelligence to do the job in the first place, but not enough for you to foresee the consequences of your choices.

You and others like you are so concerned with can you do something, that you didn't think about should you do something.

Older generations called this type of blindness and destruction evil, and viewed it as a curable malady at the turn of the century; but not a cure anyone would willingly choose.

Currently, we are on track for economic collapse by 2029, though the more they print the sooner it occurs.

Mises has a full breakdown on how centralized systems fail, written in the 1930s. Its why socialism is considered a failed economic system by rational people.


To use your phrase of: You haven't thought through the consequences of this at all. I'll break this down a bit further for you hopefully in a way you can understand.

Don't even have to go further than "You work to provide value in exchange for a store of value; currency."

Work assumes that there is work to be done. If everything is abundant, there is no work to be done because no currency needed. Who makes food? Machines. Who makes the physical infrastructure? Machines. Who generates the power? Machines.

I'm not saying that this is going to happen (I also don't think anything you're saying is going to happen either), and I think the GP was being facetious, but if you are going to take that statement seriously, then I think you do not understand what abundance truly means.


You are blind, and supporting something that is both incredibly evil and destructive.

If you can, take a look at what abundance does to people in the real world, objectively, do some actual research. Look at mortality for lottery winners, and business tycoon heirs; read their horror stories.

When there is no work, there is no purpose, no growth, no value, and no life, nothing new happens.

It is death, either quick and self-inflicted, or slow until madness from suffering takes over, where you can't notice it along the way like a person suffering from a progression of Alzheimer's. That is the abundance you seek.


I said "I'm not saying that this is going to happen" so I wouldn't say I am supporting it. Was just making a point that you looked at it too black and white and didn't think through everything. More of a response to what I felt was unnessecarry pompusness

Your argument feels more like an anarcho-primitivism/Unabomber idea of abundance or technological progress reducing humanity's satisfaction, so we should stay where we are, i.e., always have tasks(work) rather than pursue fulfilment.


That is probably not what he meant.

1. The weaknesses of the human condition might be the cause of our destruction. That's why I am not surprised when people give these anti doomer takes. History keeps repeating again and again and people haven't really learned from their mistakes.

2. The solution to this problem would be to fix or greatly diminish human greed and selfishness or anything else that I might be missing. We would then have greater chances of heading towards a utopia.

We will end up with either a utopia or a dystopia.


I don't disagree. I don't really have too much of an opinion on which one we end up with. I was more pointing out that, at the limit, it's not certain that we will end up in a dystopia through abundance, more that it is extremely uncertain of which one it will be and to extrapolate as the GP did is to not full take abundance to it's limit.


I would assume we are moving towards a dystopia. GP is correct on the dangers but his perspective is incomplete and takes you on a wrong direction as you may have said.

It would require a lot of people to understand the dangers and take action to solve the necessary problems before it's too late.

Governance and the other systems we have must fundamentally change. (All over the world)

We must acknowledge our flaws and frailty which causes us to be selfish,greedy and seek dominance and control over the rest. This is animalistic behaviour.

People must recognize that we are running out of time.

We are seeing a race whose direction is difficult to predict but the magnitude will be out of bounds.

If you are useless, you won't matter. Have you thought about how pests are treated?

This is relevant today and as time goes by, it's not difficult to predict.

I think we are heading towards a dystopian world.

I hope peace and goodwill prevails.


You sounds like those annoying religious nuts screaming the end is near and we all going to hell.


I understand what you have said but I have an argument why it can happen before 2029. How can I contact you securely? I asked this since you know the dangers, you might be extremely worried.


I'm aware of the severity but I don't have a secure way of being contacted at the moment. I don't know when I'll have it up and running.

Most of my devices have been repeatedly compromised on a fairly consistent and continuing basis (bricked at the firmware level; routers, modems, phones). HFC-head end services appears compromised on the ISP side, reports have been made; tickets opened; and closed 30 days later with no correction, and no record on future calls so its all BS.

I have been bogged down in the process of recovering from backups. Its slagged enough hardware that I now keep SPI firmware and parallel memory chip backup images for my important systems prior to connecting them to a network, and its a real chore accessing the hardware to revert to known working good images for quite a lot of these component interfaces; often they are buried destructively under pads/thermal heat sinks.

I don't have an ETA, its minute detailed work, and working directly with hardware really isn't my forte.

> you might be extremely worried.

That's been fairly consistent since this came to my attention a few years ago, not for myself but for family and friends that might survive this.

In fairness, I think its likely I'll be dead from vaccine injury by the time this rolls around. Nerve-function and muscle loss are progressive and worsening, and its stumped every expert, and worsening symptoms after reintroduction of booster, and infection later (from the grocery store, C19) so odds aren't good.

I've done what I can, and dwelling on the things you can't control doesn't do anything but create misery.

I'm aware there are a number of things that can accelerate the timetable where it would happen sooner, some outlier things as early as late 2025, but those aren't likely to happen.

2029 is the point at which the economic system fails to socialism/communism with no action needed from anyone; just business as usual. From there, its a steady death march to full collapse and most don't have a clue.


The Wu brothers burned a lot of their credibility by lying over Devin's capabilities. These are smart folks, it's disappointing they were grifting so hard.


Non-technical people are about to find out that software engineers were not kidding when they said that writing out code is actually the easy part.


I have not followed them but I got the impression that AI simply doesn’t do what they said theirs could do.


The company is still hiring software engineers so clearly the version of SWE that Devin “is” isn’t quite good enough.


man, that would suck. To build devin to replace yourself. Hope the equity is worthwhile.


It made no sense. If openai is the best, how can anything be better? Wouldn't openai done it first? If they could it should have been obvious they could make a better thing. Devin is just refined snakeoil. Bigger question is, how did they get some of you to buy it?


Most likely OpenAI would have done it first, that is assuming that LLMs are the answer. But LLMs can never be the answer since they rely on ambiguously worded free-form text written by humans. Try writing software to price a derivative that way. The only way that a computer will write accurate reliable software is if it’s given clear and precise instructions. No current tool will do that.

The answer will come from outside the usual suspects.


The idea behind Devin is an agent, folks can build better/smarter AI on top of OpenAI's models.


What do agents allow you to do that if, else, for and while statements with a database of LLM responses don't out of interest?


> What do agents allow you to do that `agents` don't out of interest?


Can you elaborate?

You mean human agents, or you are agreeing to what I'm saying here (Agents are just the same as any old backend API for practical purposes, as they can do the same things)


If you also think you described an agent yes, we agree here. There is no mystery on what's an agent, just a bunch of ifs and elses connected to a DB.


Go on?


I have noticed growing cringe on youtube shorts hyping ai products/founders 1)https://www.youtube.com/shorts/j2AwScMSrRA


I guess same as Humane Pin/Rabbit R1. Got a lot of hype then faded into obscurity.


The founder was featured in the o1 release promotional videos. Looks like he had an early access deal with OpenAI and now he's working on upgrading Devin




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: