This is why in a free/profit driven market you *absolutely* need strong regulation. A companies mandate isn't to make the world a better place. It's to make money. The government's job is there to make the world a better place (generalisation of keep balance, keep people safe, etc).
Australia gets given grief for being a nanny state, but it's what's required.
Gas HAS to be used as a backup for renewables today (or at least in places without good pumped hydro?). Enough batteries to supply an entire grid for a day or two is just out of reach today.
At least in Australia, I'm not sure how the dynamics go with such a geographically diverse grid like Europe.
Yes, certainly, I agree. But, at the same time, we should be electrifying and rapidly deploying storage (both energy and heat) as quickly as possible. "Burn the ships" and commit.
> Solar PV supplied more than 10% of global electricity in the first half of 2026, but generation remains heavily concentrated around midday. Ember says rapidly expanding, cheaper battery storage could shift 34% of new daily solar generation to non-solar hours in 2026, bringing round-the-clock solar closer.
AI is definitely eating jobs as no one would disagree with, I guess the question is more at what point does it go from low, medium, high, or apocalyptic job loss?
I don't think anyone can argue using AI to do work (rather than say using it to ask you questions or something) makes you better at the details of whatever it is that it's helping you do - maths, history, code, etc.
I think the only question is, does it matter?. Do we need to know that detail any more?
Eg AI absolutely knows fastapi middleware better than me, but I do understand what it does and why. Did I need to actually know that level of detail? I genuinely don't know the answer to that...
To be fair, some of the democratic options were fairly rubbish too (from a non-US person/perspective). I'd still have preferred you didn't pick a president that'd make my commute more expensive, and drive power/influence/money further to China/Middle East however...
Unfortunately I think the US has lost the ability to vote for a moderate/centre person or party...so it's looking pretty grim.
How much more "moderate/centrist" could the Democratic candidates be? Id'd strongly argue the issue is exactly that the US only gets to choose between a center-right and a far-right candidate.
I am not an AI evangelist by any means, but it does serve a purpose when used correctly.
I feel like if you asked any half-intelligent AI "please review these notes and flag points we should review for correctness, or check with the patient?"
I'm pretty sure it would pick up a huge chunk of issues?
> I am not an AI evangelist by any means, but it does serve a purpose when used correctly.
Well, doesn't every tool? The devil is in the details of just how easy it is to use correctly versus incorrectly, and what are the consequences if it's not used correctly. A tool that, when used as its manufacturer advises, fails as often as most AI systems do has no place in systems where peoples' health and safety can be harmed.
This is basically living in a fantasy world. Errors are routine.
One of my notes in my current medical file states I injured my shoulder playing for a NFL team. I have never played football, and certainly not at a professional level. Fixing it is sort of like trying to fix your credit report - you supposedly get it done, and then 6mo later the same error pops back up again.
Ironically the reason this note exists is very similar to a way an AI scribe would misinterpret a conversation.
My experience with whisper for example is that it tends to generate plausible-in-context content when there are long gaps or silence. I don’t think the errors as described are likely to be catchable by LLM without a lot of false positives; in context disclosing recreational drug use is something that is likely to come up in a medical consultation.
> Doctor: have you used any recreational drugs in the last six months
> Patient: No
> Doctor: < long pause as they review notes >
> Patient (hallucinated most likely response): umm, actually there was one thing I hesitated to mention. Me and my girlfriends tried microdosing…
When using an LLM for details like this I keep it out of the mode of generating prose as much and as long as possible. I get better results when I work with bullet points. I think this helps me keep the right frame of mind when reading the responses. It’s easier to scan the facts or points for accuracy. And I think it might help the LLM to be less creative with those points, but I don’t rely on that.
Two things I've figured out with pretty good certainty
1) on an existing codebase/hand written, even fable absolutely will "make a mess" if you do large changes at once, or don't check the output.
2) with detailed, WELL NUMBERED (Claude is remarkably good at following number references), multi-level design and build documentation Claude absolutely CAN produce moderately complex applications (crud with some lightly branching business workflows).
The corollary to (2) however is can you actually get that working code to be production grade, AND maintain it? that far I haven't got yet.
I have quite a bit of success with Claude on Jira tickets as long as the tasks are scoped well. Put effort into the ticket, and fire Claude at the ticket with MCP. I essentially review the PR.
But this is only possible because I have created a completely new python development workflow with fixed formatting, style, and unification of tooling around uv. We’ve also split our projects into smaller repos, some which function as libraries and some which are applications depending on those libraries.
I think the key to success here is to limit what context is required to do development. I’d say it’s slightly more on the extreme side of “modular” than a codebase I would write otherwise
Unfortunately, even with all the updates - they never fixed my main gripe - each planet is just lifeless, and it's basically the same few square kilometres copy and pasted over the whole planet (I'm about 120h played).
I wished they somehow made it more exciting to explore.
I'm also a day one NMS enthusiast. I thought I was the only one. I can definitely agree with your critique, and I think the difference was in how I played it. On release day it was basically a really pretty tech demo. You couldn't really die/fail in any meaningful way unless you tried to. But wandering the universe in a goals-agnostic sort of way, looking for different types of exocritters, learning the different alien words and doing some resource gathering/trading was satisfying enough as a second screen experience. Much like when I play American Truck Simulator, or when I focus on cargo hauling in elite dangerous, it's just kinda something to keep my hands busy while I listen to a podcast or documentary or just sit and bullshit with my wife.
Same. Played thru the original version. The overall vibe of the game was great and incredibly polished for a game half done, but yeah it was stretched very thin and started to lose feeling after a while.
That said,idk if I would like the fame at all now. Base building seems antithetical to the original nomad-like setting. And there's only loss to have adding identity to the protagonist after having played it as "yourself"
And apparently this is one of the main criticisms of Starfield. The Bethesda designers either never played No Mans Sky to understand the problem, or they simply believed a large enough budget could mask these problems.
You can do really complex and interesting stuff with procedural generation, even at-scale. You need a large breadth of well-designed rules, and then you need temporal simulation over that breadth. That simulated history is just as important as having a lot of high quality rules when you want to create a living, breathing world. Ultima Ratio Regum is an example of how important this time aspect really is, that a place needs to have a history.
The real limitation is that virtual worlds get boring. It really doesn't matter how sophisticated they are. Generative AI can't help with that. It's not a problem of the structure, or needing some additional variety, or more sophisticated set of constraints. It's the simple fact that while you can day dream about endlessly flying around a galaxy, landing on planets, eventually it just becomes rote. It's a normal biological process of adaptation and normalization. Like how when you first move somewhere, everything is fresh and full of potential, and then 3 years on everything that was novel and cool has become just background noise.
> You can do really complex and interesting stuff with procedural generation, even at-scale. You need a large breadth of well-designed rules, and then you need temporal simulation over that breadth. That simulated history is just as important as having a lot of high quality rules when you want to create a living, breathing world. Ultima Ratio Regum is an example of how important this time aspect really is, that a place needs to have a history.
The problem isn't the quality of proc-gen. The problem is that every gameplay loop roughly consists of 'scan for a thing in a proc-gen world, then click on it.'
Improving the quality of the proc-gen won't fix that.
This is a high-tech sci-fi setting! This gameplay space should have been the dev's oyster! Have resources which can be pinpointed with the scanner! Have ones that you have to triangulate the location of! Have clouds that you'd be collecting via ramscoop! Have drones following your ship! Set up a mining station! Mini-games for dealing with alien tech. More ruins and salvage opportunities than just that one species of aliens! Capture animals for questgivers! Have the actual experience of flying not be an on-rails theme park ride that's fun about twice.
I could go on and on.
This obviously costs money and time. But without it, you get a 'game' that's as wide as an ocean, and shallower than a puddle.
When I was younger I was really excited about procedural generation, because it seemed like a way to give games replayability. And it does do that for certain types of strategy and builder games, e.g. MOO2 or Factorio.
But, so far at least, it really doesn't work for first person games. You end up with a universe that's seemingly, but also endlessly boring.
There's no such thing as a limit of "procedural generation". It's a limit of the choices they made.
Look at something like DF or Rimworld - they generate pretty dense worlds with plenty of lore and things to see. Whereas Minecraft and No Man's Sky just go for physical dimensions with little content.
reply