I always saw genAI going towards NPC AI & questing (maybe in the next Bethesda or Rockstar game)... Can't say I saw this coming, and this soon. Games in 5-10 years time are going to be crazy!Improving visual fidelity I don't mind at all.
I used to play the likes of Skyrim and look for mod packs for better texture, lighting etc. This is just an out of the box way to get better visuals.
I saw an example where Harrison Ford went through DLSS 5 on Reddit
I can understand peoples grievances here but I'm pretty sure it's still experimental and NVIDIA will look at feedback, and it's also optional so it can always be turned off.
I think why not free up developers to make the gameplay as good as possible and things like DLSS5 just be something they can plug in and just work.
In some cases I can see why people are annoyed, also people are annoyed at the price of GPUs (I want an upgrade but I cant afford it) so I think there's a big reason for people who game to hate AI too.
AI and ChatGPT
- Thread starter JB08
- Start date
-
In the latest edition of the United Hour Podcast, Nik (@Rood) chats to author, Dan Williamson, about his new book 'If I Hadnt Seen Such Riches' which chronicles our 1990/91 European Cup Winners Cup victory, which is a trip down memory lane for those old enough to remember it and a vital history lesson for those who don't.
The worst one for me was their Hogwarts Legacy example. It makes a murky and slightly dirty street into something that looks like there's spotlights in every corner and completely changes the design.
Is it still worth learning Python as AI is getting more advanced?
I don't know how it's implemented, but I'd imagine they would have to be careful around likeness.
Just for the record. I would camp myself more in the AI is generally horribly used and is a bad thing camp. But, if it can squeeze more frames out of existing hardware and improve graphics here and there, then fine.
feck knows. No one who knows Python is actually writing Python outside of interviews. Instead, they just read (at best) the Python code written by LLMs.
My grim prediction, If I have to bet, is that junior and mid developers are going to have massive trouble by the end of this year (as in becoming extremely hard to get a job, and many who have jobs getting layoff), while seniors and higher will probably share the same fate by the end of the decade if not earlier.
Saying that, Python is incredibly fun cause it is very high level and you focus directly on the logic and semantics, rather than in low-level details which early on makes programming quite confusing. So my point is that it is probably good to learn for the sake of learning, rather than hoping to get money out of it.
For sure. "AI" is often wrong when churning out code, and it's good to understand why. Plus, why not? There is no need to learn about how cars work when there are people who deal with that readily available. But people still want to learn out of curiosity.
I'm in Data Engineering and my job is largely python based.
I feel more like learn how to debug, write tests, generally understand design patterns.
Python isn't hard to learn regardless but learning how to problem solve, design solutions I think is more relevant now.
I don't write a lot of code anymore but I step in when I need to.
You probably need to show proficiency for say job applications but besides that... architecting and outlining requirements I think is more relevant now.
My take is, if you're translating documents to Japanese and you don't speak the language I wouldn't hire you. Even using a LLM, better understand it than not.
I think it's probably good to learn how to code still. It's not crazy difficult and it will give you an edge...
Yeah, assuming that there will be software engineering jobs in the future, knowing how to code will be a must for such jobs, even if you are not doing it in daily basis.
And it also makes you think in a different manner, plus it’s fun. Similar to I have not solved an integral in forever, but I am glad I learned how to solve may of them.
You can't sack AI, though, can you? I can just see there being some bullshit grey area where accountability is lost in the ether and people are left hanging when things go wrong(even more so than currently).
This is my biggest worry. AI bots churning out incorrect information, coupled with a severe lack of critical thinking by those utilising it will only end one way. We're already seeing accountability delegated to AI and becoming a go to excuse for lack of oversight. Like you said, you can't fire AI. But you can blame it.
But you can fire the person blaming their mistake on AI.... Are people using this as an excuse?
https://www.theguardian.com/australia-news/2026/mar/13/ai-datacentres-environmental-impacts
But.. but you guys said no worse than playing a video game or eating meat? I dunno, I’m starting to think the experts in this thread might not actually know what they’re on about?
But.. but you guys said no worse than playing a video game or eating meat? I dunno, I’m starting to think the experts in this thread might not actually know what they’re on about?
I have no doubt about energy usage, the same for Crypto was a big issue years ago and still is. I assume that data centers will eventually help to optimize energy usage or create breakthroughs in the energy sector.
2030 is 4 years away we'll ultimately see but we can't extrapolate too far, time will tell. I guess the question is, is economic growth, curing diseases and perhaps creating the last invention we'll ever make worth the extra energy usage? Probably.
You seem to think the world is going to be taken over by AI in the very near future, but also believe that there will be 1-2-1 accountability with a human for every decision AI makes, how exactly is that going to work? Doesn't really line up, does it?
4x faster? Sounds about right everything is AI right now. Look at the markets and recent technologies...
I said if we have these companies investing money and compute into clean energy, and building out power and water facilities as they build datacenters, then we will catch up with demand and affect climate change in a neutral to positive way. Let's make that happen as humans.
Yes I believe AI will move into most things like electricity, and that they're needs to be human accountability, for trust... Which is going to be increasingly important. Perspectives are nuanced right? How do you see a business or a process with no human invention in the near future playing out? Sounds silly to me...
I am currently building an agent, I want to be able to authorise any 'dangerous' or meaningful action it should want to take, and I'll create guard rails to ensure that. Anything else is crazy naive if you understand these things.
Last edited:
Focus on general CS. Architecture, Programming Patterns, maybe focus on scalable systems later on, where AI is really shit in my own experience. Or move closer to the metal, robotics, hardware engineering, where AI is also pretty shit (I hear, no experience myself). You can of course choose Python as your starting language. However, just "Learning Python" might make you a proficient Python Coder, but Coding might indeed by "going away" soon-ish. Software Engineering, on the other, is not going anywhere.
Leverage AI to learn but not to solve every problem you encounter. That's how you build "taste", which no AI has, they just pattern match on ~ median stackoverflow slop.
Pretty much this entire past year AI is now fully integrated into all my creative workflows. Been using for vfx, mixed realities, design, programming, previs, 3d animation and rendering. Don't think there's a single project that I do now that doesn't involve it. Quite surprised at how quick I've not only adapted to it but how the entire creative industry is now in its thrall.
Thank you.
Same as now - with the company providing the service/product. I don't get the problem here. It's not like Google is going to take responsibility if you fecked up the research. Individuals are never taking the blame anyway.
Nvidia DLSS5 announcement..... What the feck is this shit.
I genuinely thought it was a joke but it isn't. Digital Foundry must have been paid for their analysis of it because they think it looks amazing. It just looks like AI slop.
Nvidia played this marketing wrong. What the feck is this video, half of it is a recording of a TV screen on the wall.
This is the right use of the technology that can make a revolution in the gaming industry. Not sure what is the complaining about. It's up to the user if he wants to enable this. Looks pretty good imo.
The answer depends where are you now. I think if you are already around IT (like Business Analysts, Product Owners, Technical Project Manager etc) it makes perfect sense to learn Python as your main programming language, even just for understanding how coding works, and it can be useful in scripts/data transformation. AI will then increase /simplify/ speed up usage of Python for you.
If you are doing something completely different, or are aiming to become a developer, saying "I know Python" on your CV is an equivalent of "I speak English". It doesn't give you much advantage. And already going to second question you might ask: I don't know what language makes sense to learn now. I run a lot of interviews for software developers in different technology and there's a lot of candidates on the market already (some of them lost jobs, others are looking for a change because don't feel secure).
It seems the IT train has left the station. There's more supply than demand, at least for developers. I think it's easier to get into IT for different jobs, but they require some specific business knowledge/processes. I know a lot of people who got into IT that way.
Last edited:
It is indeed used for mostly meaningless and harmful shite.
This is a pretty cool idea.
Andrei Karpathy created this project Autoresearch to train, evaluate and optimise a mini version of GPT2. He provided the scaffolding in a single python file and after that just passed it over to the AI agent with instructions in a text file, and it ran dozens of experiments overnight to make the model more efficient and powerful. It's quite a narrowly scoped task, but that task occupied a significant part of many highly-paid ML engineers' jobs not long ago: train a big neural network, track its performance as it trains, develop hypotheses and test different hyperparameters to improve performance, and loop into you reach a performance target or time limit.
Lots of people have built spin-off ideas from that:
- AutoResearchClaw connects it to OpenClaw and runs it through a full research pipeline of literature review, hypothesis testing, analysis and report writing.
- autoresearch@home sets up a space for different people's agents to collaborate together in a "swarm", each one proposing their own optimisation and learning from each other
So long as there’s any humans left in the mix at all, the buck will stop with them. Even if that’s only the board of directors. Although they’re bound to keep a layer or two of shmucks beneath them to be fall guys when necessary.
I don't see how it works with a non-deterministic framework, it will never be possible to completely rely on it, yet nobody is going to want to keep enough humans around to validate everything, and there will be less and less humans that you can trust to validate as nobody is being hired as a junior to eventually end up being experienced enough to do so.
It's not that I don't think what you and others are saying is the likely setup, I just don't see how it will work well for the end users, it's going to end up with less accountability and more opportunity for screwing the consumer.
I’m definitely not arguing it will work well for the end user. Just saying the buck will always stop with a human. Which I actually think is going to be a massive headache for anyone working for a business that relies on AI for important processes/decisions (basically all businesses in a few years time?)
It doesn't matter what they may want to do.... Clients/customers are going to want something they can trust. That will dictate the amount of oversight that they need to have in place, that they can live with. Otherwise people will go elsewhere. Like I said trust is going to be increasingly more important.
If you look on moltbook it's said that the AIs don't even trust each other, so why would we trust them to run a whole company for example with no oversight?
True but I guess it's a job created to go along with all the jobs we are about to lose, so we don't all need to become plumber yet? I'm kidding......
This is my exact point, and what I'm seeing in my industry is that it pretty much sounds like companies are going to sacrifice more feck ups for the upside in cost savings/productivity. Which is not ideal in a lot of scenarios.
You can severely curtail an AI agent to work within a framework though and test the shit out of it. That's what I'm seeing in my industry and then obviously you have operational control reports on top and you have a secondary QC (sometimes 1 AI QC and another human QC dip test) in the workflow.
At some stage it'll all become a bit more stable and it'll be considered no different to RPA. At the moment with the pace of change firms will be using it as a shortcut due to it's general capabilities and I expect many large feck ups due to poor engineering and testing (if there's any human testers left anywhere!).
The current business assumption is a bit less quality but a lot more savings. Gradually, service levels have dropped because you first need to go through an AI agent. Worse customer experience? Yes. Are you going to go to competition? No, because they are doing the same.
Regarding "less and less humans that you can trust to validate" - I don't think this is going to be the case. Even now business/IT processes include "4 eye principle" for validation/testing. You will need more skilled people to do that job yes, but if the job market continues to shrink like it does now, plenty of people will be happy to do that job.
It's similar to self-service checkout. You put 4 of those, you fire 2 people, and upskill the remaining ones to handle validation/special cases.
Is the customer more happy that he is doing the checkout himself? Maybe. Mostly only because he does not stand in a queue. Vast majority still prefers the traditional way.
Is the employee happy he needs to jump between the kiosks and solve problems? Very much doubt that. Are they leaving the job? Also doubt that.
Who wins? Owners/corporate. More savings, better results.
This. 100% this.
Thanks to enshittification the customer has been sucking up a steady decline in the experience of using basically all tech based products for years now. Nobody is going to (or will be able to?) walk away if that decline continues.
In an age where you can vibe code your own saas, and your own apps. When you can build and run your own agent to your own specification. As a consumer, as a business, whatever... if a service or product is making mistakes and not fitting your requirements, blaming it on AI and lack of oversight... in that scenario if you choose to stick with them then it's on you. It's not like you don't or won't have choice.
Next time my 78 year old dad is in the weeds, dealing with garbage customer service from a big corporation, I should tell him it’s his fault for not vibe coding an app or training an AI agent to do it for him?
Bro, I'm trying to tell to you shop around but you seem fixated on what I said and how I said it rather than the possibility that I'm right? It's hard work in here.... Sometimes I say hard truths, it's like people don't want to hear, and in those cases it's best to ignore me. Do we want everyone singing from the same hymn sheet all the time? Shop around, speak to people in the field, pay a mate like you would with other stuff, etc.... don't just accept the fukry?
I'm assuming you'll help him when the time comes right?
Last edited:
Complete and utter nonsense as usual, there will be a plethora of services that people will have to use for some reason or another, just wait until the NHS becomes an agentic wasteland because your incompetent government allows it to happen for a few brown envelopes and see how you're feeling about it.
Maybe it will, but I'm speaking about your original point and Pogue's addendum. If a company is making mistakes and blaming it on AI it is an individuals choice whether to stay or not. If it's AI services there are many competitors as well as a rise in hobbyists, if its utilities then we are already encouraged to switch even though many don't, and if its a service like NHS or Transport where we are choice limited, then you have to force the change you want to see, not accept what you are given. That's all I'm saying and I don't feel like adding anymore. I get accused of not engaging even though I often do, but I'm kinda over circular arguments and I realise this subject is emotive and divisive.
I feel the friction between many of us is I just don't vibe with pessimistic "things are going to shit and we cant do anything about it" thinking... and that's ok, as is the pessimistic thinking, we are all different, it is what it is...
Last edited:
My initial point was about the services I'm talking about, important things that have consequences if there's a degradation in quality and/or accountability. These comparisons to Tesco self checkout or some meaningless optional services are completely tangential to what I've been talking about.
We’re discussing a deterioration in the quality of services provided to us punters as corporations use AI to cut costs and corners. Implying this is on the customers for not using AI to shop around for alternative providers is a terrible take. We all know that a lot of these providers are either running what is pretty close to a monopoly or they’re all in it together. So shopping around is often not an option.
I know your whole vibe in this thread is “educate yourself, the change is coming regardless and you need to be ready” and that’s fine but there’s a whole world out there of people who are barely coping with the shift from real world to online services. So it’s ridiculous to imply that anyone who gets screwed by a corporation’s crappy implementation of AI should feel it’s their own fault for not using agentic AI to solve all their problems (or whatever you’re actually suggesting here)
I didn't say to use AI to shop around, I didn't say to vibe code an app or train an AI agent (thats something that I can and will do and it's probably a good idea to have someone like that in your circle). I also didn't compare to self checkout. There's a whole lot of talking passed each other, and assumptions, from all of us. And it's an emotive subject which doesn't help because sometimes emotion gets in the way of logic, for all of us.
Who said that? I specifically said, if you accept the service, then its on you.
Can I ask you... do you think we should be preparing and informing people in as balanced a way as we can? Should we be trying to affect change now? Or accepting what comes (including the worst case scenarios), waiting to see how things pan out?
And on a side note, I already know that not everyone in this thread thinks I talk shite, so maybe the ones that do should just do themselves a favour and pop me on ignore?
Last edited:
This is a good video. Essentially, AI makes you delusional in your own abilities and think you’re way better than you are. It’s a pattern I’ve noticed, the engineers and even non engineers I’ve seen who’ve basically succumbed to Claude to do all their work for them have turned into the worst yet most arrogant engineers.