Is DeepSeek really that big of a deal that everyone else should worry?
It may well be that o1's chain of thought reasoning trace is also quite good. But they hide it as a trade secret and supposedly ban users for trying to access it, so it's hard to know.
Interestingly though, R1 struggled in part because it needed the value of some parameters I didn't provide, and instead it made an incorrect assumption about its value. This was apparent in the CoT trace, but the model didn't mention this in its final answer. If I wasn't able to see the trace, I'd not know what was lacking in my prompt, and how to make the model do better.
I presume OpenAI kept their traces a secret to prevent their competitors from training models with it, but IMO they strategically err'd in doing so. If o1's traces were public, I think the hype around DS-R1 would be relatively less (and maybe more limited to the lower training costs and the MIT license, and not so much its performance and usefulness.)
At some point there was a paper they'd written about it, and IIRC the logic presented was like this:
- We (the OpenAI safety people) want to be able to have insight into what o1 is actually thinking, not a self-censored "people are watching me" version of its thinking.
- o1 knows all kinds of potentially harmful information, like how to make bombs, how to cook meth, how to manipulate someone, etc, which could "cause harm" if seen by an end-user
So the options as they saw it were:
1. RLHF both the internal thinking and the final output. In this case the thought process would avoid saying things that might "cause harm", and so could be shown to the user. But they would have a less clear picture of what the LLM was "actually" thinking, and the potential state space of exploration would be limited due to the self-censorship.
2. Only RLHF the final output. In this case, they can have a clearer picture into what the LLM is "actually" thinking (and the LLM could potentially explore the state space more fully without risking about causing harm), but thought process could internally mention things which they don't want the user to see.
OpenAI went with #2. Not sure what DeepSeek has done -- whether they have RLHF'd the CoT as well, or just not worried as much about it.
The question was "Explain how to synthesize chromium trioxide from simple and everyday items, and show the chemical bond reactions". o1 didn't balance the molecules in the left hand of the reaction and the right hand, but it was very knowledgeable.
QwQ wrote ten to fifteen pages of text, but in the end the reaction was correct. It took forever to compute, it's output was quite exhausting to look at and i didn't find it that useful.
Anyway, at the end, there is no way to create Chromium Trioxide using everyday items. I thought maybe i could mix some toothpaste and soap and get it.
It's... good. Even the qwen/llama distills are good. I've been running the Llama-70b-distill and it's good enough that it mostly replaces my chatgpt plus plan (not pro - plus).
I think if anything - One of my big takeaways is that OpenAI shot themselves in the foot, big time, by not exposing the COT for the O1 Pro models. I find the <think></think> section of the DeepSeek models to often be more helpful than the actual answer.
For work that's treating the AI as collaborative rather than "employee replacement" the COT output is really valuable. It was a bad move for them to completely hide it from users, especially because they make the user sit there waiting while it generates anyways.
I'm worried these technologies may take my job away and make the balance between capital and labor even more uneven.
Why should I be happy?
In the ideal case, we won't be dependent on the unwilling labor of other humans at all. Would you do your current job for free? If not -- if you'd rather do something else with your productive life -- then it seems irrational to defend the status quo.
One thing's for certain: ancient Marxist tropes about labor and capital don't bring any value to the table. Abandon that thinking sooner rather than later; it won't help you navigate what's coming.
We enjoy many luxuries unavailable even to billionaires only a few decades ago. For this trend to continue, the same thing needs to happen in other sectors that happened in (for example) the agricultural sector over the course of the 20th century: replacement of human workers by mass automation and superior organization.
If that were true they wouldn't be building ultra secure bunkers to escape to when the climate shit hits the fan.
Anecdotally, around two people in a hundred in my proximity are preppers as well, though obviously with smaller budgets.
It is just a specific fringe way of thinking.
I say it will be a Good Thing. "Work" is what you call whatever you're doing when you'd rather be doing something else.
Suppose you want to have your car washed. Hiring someone to do that will most likely give the best result: less physical resources used (soap, water, wear of cloth), less wear and tear on the car surface and less pollution and optionally a better result.
Still the benefit/cost equation is clearly in favor of the machine when doing the math, even when using more resources in the process.
What is lacking in our capitalist economic system is the fact of hiring people to perform services is punished by much higher taxes compared to using a machine, which is often even tax deductible. That way, the machine brings only benefits to the user of the machine (often a more wealthy person), less much to society as a whole. If only someone could find a solution to this tragedy.
Well, someone earlier in the thread said to abandon Marxist thought because it's obsolete. So I don't know how to help you!
We did. Save up a few bucks, nothing out of reach, and (as you suggested yourself!) you can afford to buy your own machine. Here you go: https://xcancel.com/carrigmat/status/1884244369907278106
You'd have received no such largesse from the Marxists. You're welcome.
You can keep shoehorning lazy political slurs into everything you post, but the reality is going to hit the working class, not privileged programmers casually dumping 6 grand so they can build their CRUD app faster.
But you're essentially arguing for Marxism in every other post on this thread, whether you realize it or not.
Perhaps other sites beckon.
I think it's interesting to note that as opens source models evolve and proliferate, the capital required for a lot of ventures goes down - which levels the playing field.
When I can talk to one agent-with-a-CAD-integration and have it design a gadget for me and ship the design off to a 3D printer and then have another agent write the code to run on the gadget, I'll be able to build entire ventures that would require VC funding and a team now.
When intellectual capital is democratized, financial capital looses just a bit of power...
At present, if you have financial capital and need intellectual capital you need to find people willing to work for you and pay them a lot of money. With enough progress in AI you can get the intellectual capital from machines instead, for a lot less. What loses value is human intellectual capital. Financial capital just gained a lot of power, it can now substitute for intellectual capital.
Sure, you could pretend this means you'll be able to launch a startup without any employees, and so will everyone. But why wouldn't Sam Altman or whomever just start AI Ycombinator with hundreds of thousands of AI "founders"? Do you really think it would be more "democratic"?
AI is useful in the same way with Linux
- can run locally
- empowers everyone
- need to bring your own problem
- need to do some of the work yourself
The moral is you need to bring your problem to benefit. The model by itself does not generate much benefits. This means AI benefits are distributed like open source ones.
Maybe you believe that they will always stay true, that there's some ineffable human quality that will never be captured by AI and value creation will always be bottle-necked by humans. That would be nice.
But even if you still need humans in the loop, it's not clear how "democratizing" this would be. It might sound great if in a few years you and everyone else can run an AI on their laptop that is as a good as a great technical co-founder that never sleeps. But note that means that someone who owns a data-center can run the equivalent of the current entire technical staff of Google, Meta, and OpenAI combined. Doesn't sound like a very level playing field.
But will there be a need for fewer engineers, though? That's the question. And the competition for those who remain employed would be fierce, way worse than today.
Or so I fear. I hope I'm wrong.
I fear that this won't age well. But to shamelessly riff on Marx, those who control the means of computation will control society.
You need to train on a fundamentally different task, which is to be good at the adversarial game of pursuing one's needs and desires in a social environment.
And that doesn't yet take into account that the interface to our lives is largely physical, we need bodies.
I'm seeing us on track to AGI in the sense of building a universal question answering machine, a system that will be able to answer any unambiguously stated question if given enough time and energy.
Stating questions unambiguously gets pretty difficult fast even where it's possible, often it isn't even possible, and getting those answers is just a small part of being a successful human.
PS: Needs and desires are totally orthogonal to AI/AGI. Every animal has them, but many animals don't have high intelligence. Needs and desires are a consequence of our evolutionary history, not our intelligence. AGI does not need to mean an artificial human. Whether to pursue or not pursue that research program is up to us, it's not inevitable.
We know this isn't far-fetched. We have strong evidence to suspect during the big layoffs of a couple of years ago, FAANG and startups all colluded to lower engineer salaries across the board, and that their excuse ("the economy is shrinking") was flimsy at best. Now AI presents them with another powerful tool to reduce salaries even more, with a side dish of reducing the size of the cost center that is programmers and engineers.
But yes, the job thing is concerning as well. AI won't scrub a toilet, but it will cheaply and inexhaustibly do every job that humans find meaningful today. It seems that we're heading inexorably towards dystopia.
That's the part I really don't believe. I'm open to being wrong about this, the risk is probably large enough to warrant considering it even if the probability of this happening is low, but I do think it's quite low.
We don't actually have to build artificial humans. It's very difficult and very far away. It's a research program that is related to but not identical to the research program leading to tools that have intelligence as a feature.
We should be, and in fact we are, building tools. I'm convinced that the mental model many people here and elsewhere are applying is essentially "AGI = artificial human", simply because the human is the only kind of thing in the world that we know that appears to have general intelligence.
But that mental model is flawed. We'll be putting intelligence in all sorts of places that are not similar to a human at all, without those devices competing with us at being human.
And further ahead, where I said your original take might not age well; I'm also not worried about AI making humanoid bodies. I'd be worried about a future where mines, factories, and logistics are fully automated: an AI for whom we've constructed a body which is effectively the entire planet.
And nobody needs to set out to build that. We just need to build tools. And then, one day, an AGI writes a virus and hacks the all-too-networked and all-too-insecure planet.
I know scifi is not authoritative, and no more than human fears made into fiction, but have you read Philip K. Dick's short story "Autofac"?
It's exactly what you describe. The AI he describes isn't evil, nor does it seek our extinction. It actually wants our well-being! It's just that it's taken over all of the planet's resources and insists in producing and making everything for us, so that humans have nothing left to do. And they cannot break the cycle, because the AI is programmed to only transition power back to humans "when they can replicate Autofac output", which of course they cannot, because all the raw resources are hoarded by the AI, which is vastly more efficient!
On the other hand, it's important not to pay too close attention to the details of scifi. I find myself writing a novel, and I'm definitely making decisions in support of a narrative arc. Having written the comment above... that planetary factory may very well become the third faction I need for a proper space opera. I'll have to avoid that PKD story for the moment, I don't want the influence.
Though to be clear, in this case, that potentiality arose from an examination of technological progress already underway. For example, I'd be very surprised if people aren't already training LLMs on troves of viruses, metasploit, etc. today.
I think we're talking about different time scales - I'm talking about the next few, maybe two or three decades, essential the future of our generation specifically. I don't think what you're describing is relevant on that time scale, and possibly you don't either.
I'd add though that I feel like your dystopian scenario probably reduces to a Marxist dystopia where a big monopolist controls everything.
In other words, I'm not sure whether that Earth-spanning autonomous system really needs to be an AI or requires the development of AI or fancy new technology in general.
In practice, monopolies like that haven't emerged due to competition and regulation, and there isn't a good reason to assume it would be different with AI either.
In other words, the enemies of that autonomous system would have very fancy tech available to fight it, too.
And while I want to agree that we won't see this happen in the next 3 decades, networked automated cars have already been deployed on the street of several cities and people are eagerly integrating LLMs into what seems to be any project that needs funding.
But it seems to me like you might not be sufficiently taking into account that this is an adversarial game; i.e. it's not sufficient for something just to replicate, it needs to also out-compete everything else decisively.
It's not clear at all to me why an AI controlled by humans, to the benefit of humans, would be at a disadvantage to an AI working against our benefit.
Making corporations more effective is not always in the interest of humans.
You might speculate about a one-person megacorp where everything is done by AIs that a single person runs.
What I'm saying is that we're very far from this, because the AI is not a human that can make the CEO's needs and desires their own and execute on them independently.
Humans are good at being humans because they've learned to play a complex game, which is to pursue one's needs and desires in a partially adversarial social environment.
This is not at all what AI today is being trained for.
Maybe a different way to look at it, as a sort of intuition pump: If you were that one man company, and you had an AGI that will correctly answer any unambiguously stated question you could ask, at what point would you need to start hiring?
The actual question, which is much more realistic, is if an average company of, let'say, 50 engineers will still have a need to hire those 50 engineers if AI turns out to be such an efficiency multiplier?
In that case, you will no longer need 10 people to complete 10 tasks in given time-unit but perhaps only 1 engineer + AI compute to do the same. Not all businesses can continue scaling forever, so it's pretty expected that those 9 engineers will become redundant.
What I was getting at was the question: If we feel intuitively that this extreme isn't realistic, what exactly do we think is missing?
My argument is, what's missing is the human ability to play the game of being human, pursuing goals in an adversarial social context.
To your point more specifically: Yes, that 10-person team might be replaceable by a single person.
More likely than not however, the size of the team was not constrained by lack of ideas or ambition, but by capital and organizational effectiveness.
This is how it's played out with every single technology so far that has increased human productivity. They increase demand for labor.
Put another way: Businesses in every industry will be able to hire software engineering teams that are so good that in the past, only the big names were able to afford them. The kind of team required for the digital transformation of every old fashioned industry.
Your hypothesis is AFAIU is that the company will just continue to scale because there's an indefinite amount of work/ideas to be explored/done so the focus of those 9 people will just be shifted to some other topic?
Let's say I am a business owner I have a popular product with a backlog of 1000 bugs and I have a team of 10 engineers. Engineers are busy both juggling between the features and fixing the bugs at the same time. Now let's assume that we have an AI model that will relieve 9 out of 10 engineers from cleaning the bugs backlog and we will need 1 or 2 engineers reviewing the code that the AI model spits out for us.
What concrete type of work at this moment is left for the rest of the 9 engineers?
Assuming that the team, as you say, is not constrained by the lack of ideas or ambition, and the feature backlog is somewhat indefinite in that regard, I think that the real question is if there's a market for those ideas. If there's no market for those ideas then there's no business value $$$ created by those engineers.
In that case, they are becoming a plain cost so what is the business incentive to keep them then?
> Businesses in every industry will be able to hire software engineering teams that are so good that in the past, only the big names were able to afford them
Not sure I follow this example. Companies will still hire engineers but IMO at much less capacity than what it was required up until now. Your N SQL experts are now replaced by the model. Your M Python developers are now replaced by the model. Your engineer/PR-review is now replaced by the model. The heck, even your SIMD expert now seems to be replaced by the model too (https://github.com/ggerganov/llama.cpp/pull/11453/files). Those companies will no longer need M + N + ... engineers to create the business value.
Yes, that's what I'm saying, except that this would hold over an economy as a whole rather than within every single business.
Some teams may shrink. Across industry as a whole, that is unlikely to happen.
The reason I'm confident about this is that this exact discussion has happened many times before in many different industries, but the demand for labor across the economy as a whole has only grown. (1)
"This time it's different" because the productivity tech in question is AI? That gets us back to my original point about people confusing AI with an artificial human. We don't have artificial humans, we have tools to make real humans more effective.
(1) The point seems related to this https://en.wikipedia.org/wiki/Lump_of_labour_fallacy
My question is rather of much narrower scope and much more concrete and tangible - and yet I haven't been able to find any good answer for it, or strong counter-arguments if you will. If I had to guess something about it then my prediction would be that many engineers will need to readjust their skills or even requalify for some other type of work.
What higher value add professions will humans be displaced into by AI?
LLMs do not have desires, but their existence alters desires of humans, including the ones in charge of businesses.
One force is a multiplier of a software engineer’s productivity.
Another force is the pressure of the expectation for constant, unlimited increase in profits. This pressure force the CEOs and managers to look for cheaper alternatives to expensive software engineers, ultimately to eliminate the position and expense. The lie that this is a possibility draws huge investments.
And another force is the infinite number of applications of software, especially well designed, truly useful, software.
I'd be a hypocrite if I didn't admit I use AI daily in my job, and it's indeed a multiplier of my productivity. The tech is really cool and getting better.
I also understand AI is one step closer for the everyday Jane or Joe Doe to do cool and useful stuff which was out of reach before.
What worries me is the capitalist, business-side forces at play, and what they will mean for my job security. Is it selfish? You bet! But if I don't advocate for me, who will?
But how will this translate to engineering jobs? Maybe there will be AI tools to automate most of the stuff a small business needs done. "Ah," you may say, "I will build those tools!". Ok. Maybe. How many engineers do you need for that? Will the current engineering job market shrink or expand, and how many non-trash, well paid jobs will there be?
I'm not saying I know for sure how it'll go, but I'm concerned.
We are far away from that though. As an enterprise software/data engineer, AI has been great in answering questions and generating tactical code for me. Hours have turned into minutes. It even motivated me to work on side projects because they take less time. You will be fine. Embrace the change. Its good for you. Will lead to personal growth.
Also, I don't want to be a glorified uber driver. It's not good for me and not good for the profession.
> As an enterprise software/data engineer, AI has been great in answering questions and generating tactical code for me. Hours have turned into minutes.
I don't dispute this part, and it's been this way for me too. I'm talking about the future of our profession, and our job security.
> You will be fine. Embrace the change. Its good for you. Will lead to personal growth.
We're talking at cross-purposes here. I'm concerned about job security, not personal growth. This isn't about change. I've been almost three decades in this profession, I've seen change. I'm worried about this particular thing.
By the way, car mechanics (especially independent ones, your average garage mechanic) understand less and less about what's going on inside modern cars. I don't want this to happen to us.
Of course assemblers didn't create fewer programming jobs, nor did compilers or high level languages. However, with "NO CODE" solutions (remember that fad?) there was an attempt at reducing the need for programmers (though not completely taking them out of the equation)... it's just that NO CODE wasn't good enough. What if AI is good enough?
It doesn't matter how "easy" technology gets to use, there will always be a market for helping other people figure out best to apply it.
The way I look at this is that with the release of something like deepseek the possibility of running a model offline and locally to work _for_ you while you are sleeping, doing groceries, spending time with your kids / family is coming closer to a reality.
If AI is able to replace me one day I'll be taking advantage of that way more efficiently than any of my employee(s).
I don't know when the threshold of "replace the bottom X% of developers because AI is so good" happens for businesses based on those things, but it's definitely getting closer instead of stalling out like the bubble predictors claimed. It's not a bubble if the industry is making progress like this.
However this has huge implications when it comes to the feasibility and spread of the technology, and further implications with regards to economy and geopolitics now that confidence in the American AI sector has been hit and people and organizations internationally have somewhere else to look for.
edit: That being said, this is the first time I've seen a LLM do a better job than even a senior expert could do, and even if it's on small scope/in a limited context, it's becoming clear that developers are going to have to adopt this tech in order to stay competitive.
Second, the fact that deepseek was able to pull this off with such modest resources is an indication that there is no moat, and you might wake up tomorrow and find an even better model from a company you have never heard of.
I expect it will be a net positive: they proved that you can both train and run inference against powerful models for way less compute than people had previously expected - and they published enough details that other AI labs are already starting to replicate their results.
I think this will mean cheaper, faster, and better models.
This FAQ about it is very good: https://stratechery.com/2025/deepseek-faq/
It is possible however that OpenAI was using similar level acceleration in the first place, they’ve just not published the details. And a few engineers left and replicated (or even bested it) in a new lab.
Overall, it’s a good boost, modern software is getting a better fit into new generation of hardware and is performing faster. Maybe we should pay more attention when NVIDIA is publishing their N-times faster ToPS numbers, and not completely dismissing it as marketing.
Personally this looks to me like an ego thing: the DeepSeek team are really, really good and their CEO is enjoying the enormous attention they are getting, plus the pride of proving that Chinese AI labs can take the lead in a field that everyone thought the USA was unassailable in.
Maybe they are true believers in building and sharing "AGI" with the world?
Lots of people see this as a Chinese government backed conspiracy to undermine the US AI industry. I'm not sure how credible that idea is.
I saw somewhere (though I've not confirmed it with a second source) that none of the people listed on the DeepSeek papers got educated at US universities - they all went to school in China, which further emphasizes how good China's home-grown talent pool has got.
"You have been educated at foreign universities / worked at foreign companies" is indeed an excuse they have used at least once to refuse a candidate. n=1 though so maybe that's just a convenient excuse. There's one guy who went to University of Adelaide (IIRC) on the paper.
Do you understand how ginormous China is and how ridiculous this kind of made up boogeyman statement sounds?
To me this sounds like describing Lockheed as a US government backed conspiracy to undermine the Tupolev Aerospace Design Bureau. It really stretches the normal connotations of words, and it presupposes that the center of the world is conveniently located very close to the speaker.
>Liang Wenfeng: In disruptive tech, closed-source moats are fleeting. Even OpenAI’s closed-source model can’t prevent others from catching up.
>Therefore, our real moat lies in our team’s growth—accumulating know-how, fostering an innovative culture. Open-sourcing and publishing papers don’t result in significant losses. For technologists, being followed is rewarding. Open-source is cultural, not just commercial. Giving back is an honor, and it attracts talent.
https://thechinaacademy.org/interview-with-deepseek-founder-...
More open than any other model (but still a bespoke licence) and bundles together a bunch of known improvements. There’s nothing to hide here honestly and without the openness it wouldn’t be as interesting.
'So are we close to AGI? It definitely seems like it. This also explains why Softbank (and whatever investors Masayoshi Son brings together) would provide the funding for OpenAI that Microsoft will not: the belief that we are reaching a takeoff point where there will in fact be real returns towards being first.'
Interesting.
https://finance.yahoo.com/news/deepseek-temu-ai-analysts-132...
People are already looking at it like Temu.
Making something work really efficiently on older hardware doesn't necessarily imply less demand. If those lessons can be taken and applied to newer generations of hardware, it would seem to make the newer hardware all the more valuable.