It's interesting how genre coding have become much more enforced these years. Back in the early 2010s we would call most of these just "sci-fi", because the influence and aesthetics of cyberpunk has already bled into the wider genre that you'd be more hard pressed to find a story without some influence on it.
But despite the author's veneration of 90s and 2000s, I'd argue the works from the late 2000s and early 2010s are just as sophisticated and arguably more relevant to our modern world; Project Itoh (Genocidal Organ, Harmony), Gatchaman Crowds, Beatless, alot of their themes were prescient; Because the frontier isn't hardware, it's ontology and semantics.
This is wrong. Ma was put down because of a speech he made attacking the banking system as outdated and needing reform. It was the P2P lending given that the whole thing was the government's own initiative from the late Li Keqiang and they approved the IPO right till the speech.
This not exactly wrong, but also not right / poor timeline & PRC domestics politics reading.
LKQ was pushing P2P lending / light regulatory on internet finance in ~2015.
Ant group exploited light guidance into basically shadow banking with systemic risk over next few years. PBOC had to step in to fix bad LKQ guidance.
PBOC issued rules regulating P2P lending loopholes one month before Ant Group IPO specifically calling out Ant Group. Anyone not retarded knew this means Ant Group must reform for smooth IPO, i.e. politically securities watchdog approval was going to be predicated on PBOC instructions being taken seriously. Then Jack Ma did a full retard and tried to challenge PBOC, so IPO blocked.
Well 50% retarded because ANT record breaking 300B IPO was predicated on Ant continuing to exploit low leverage shadow banking that socialized loss to state banks - hence PBOC mandated internet finance P2P to fund 30% of loans vs 2% ANT was getting away with, which would have tanked IPO.
The mismanagement here was actually Microsoft giving studios free reign and they flopped with projects like Starfield. This needs to be understood really within the larger industry trend of quality decline of AAA, which I suspect has to do with changes in overall dev culture and discourse than corporate decisions. There hasn't been a time when developers have been more disconnected from their audience than now.
> There hasn't been a time when developers have been more disconnected from their audience than now.
There are studios that routinely put out games that people love. For example, my wife and I will happily pre-order the next co-op game from Hazelight Studios. FromSoftware, Ghost Ship Games, and others fall into the same boat for many people.
The difference, in my opinion, is that games from these studios don't focus on creating an "always on" service, microtransactions, day-1 DLC, MRR, etc. Those things that blight games made by corporate studios are evidence of corporate executives putting their thumb on the scale.
There's only so much you can do, as a developer, to polish a turd.
With the obvious exception of Elden Ring, most of the games from these studios are quite modest small to medium sized projects.
I wish devs would stop trying to make every game an uber-game. They need all the monetization because they've already blown the budget before work even started.
Nintendo figured this out. When will the other big players?
At one point you could have said the same thing about Blizzard, that everything they did was gold. It’s hard to pinpoint where they went wrong, but it’s not clear to me that it was business model or game format related. Like, WoW was a fabulous success from day 1 and it was a live service game.
For games like Starfield, monetization isn't even the main issue but rather political ideology. You can only do so much ESG-approved preaching before customers go elsewhere.
> There hasn't been a time when developers have been more disconnected from their audience than now.
I'm curious what you mean by this. It seems like gamers have never been more vocal and there have never been more avenues (social media, short form video platforms, etc.) for them to voice their opinions than we have now.
How exactly do you know that developers have never been more disconnected from their audience? And how would that be relevant to declining AAA game quality when it's the responsibility of management and leadership to ensure the quality of the final product?
I think developers = studio in this case, management included. It’s the same issue as Hollywood really, making content for a loud minority will not generate revenue.
Compare how Marvel Rival performs to any Marvel Hollywood content after Endgame.
One way is developers making game for far left silicon valley ideologes when their audience is full of horny young males. I.e. the game is primarily designed to push an agenda rather than be fun.
This is a trope that's more than a few years out of date. Silicon Valley is run by Mark Zuckerberg, Peter Thiel, Marc Andreessen, Elon Musk, and many other executives who are completely devoted to the current administration; even Tim Cook bent the knee and donated $1 million to Trump's inauguration.
And what does "far left" mean in this context? Is it the abolition of private property and the creation of a communist utopia? Or are we talking about basic equal rights and a slightly more progressive tax system?
Without a bit more clarity, comments like yours come off as being uninformed at best, ignorant at worst.
Free rein is bad management by definition. You don’t need a manager to have free rein.
I’d be curious what you think the changes in dev culture are. I have worked for or with a lot of these studios and to me they have different cultures. But I could be missing the forest for the trees. MSFT has one culture that imo lacks a creative vision.
CNNs are not SoTA anymore when it comes to large models, and also are not used to provide interpretations of images as text, but rather to classify, do semantic segmentation, etc.
CNNs are fine when trained with a good recipe. There are very few good studies comparing them with proper hyperparam search and all the training tricks applied consistently. Transformers are good but ViT vs CNN is not some settled issue. Transformers are more hyped and more popular with the tech enthusiasts who just read forums and news, but if you need stuff done, CNNs are still great.
I agree, but since we're talking about imagine understanding with text output, clearly a CNN is unsuitable. My previous comment was overly reductive and CNNs can still be SoTA depending on your performance metrics. I spent the earlier part of my career training CNNs, and they are very pleasant to work with.
>Transformers are more hyped and more popular with the tech enthusiasts who just read forums and news, but if you need stuff done, CNNs are still great.
Vits are straight up more popular for ML research now, it's not just 'tech enthusiasts'.
Yes but not based on rigorous comparison. I'm not saying ViT is bad. But it took over mainly because it's the shiny new thing. It very bandwagon-Y even among PhD students.
> There's no 'rigorous comparison' that puts CNNs over Vits
That’s not accurate. My team wrote a paper for school in which a resnet model out performed a ViT model of the same size on almost all metrics. These were smaller models, but depending on the use case that might be what you want.
Don't know if it's you (did you publish?). I read about something similar but it had its issies:
- Tuning hyperparameters to gain improvement on a dataset when you're constantly looking at the answers is pretty meaningless. It's basically testing on the training data.
- Eval on ImageNet1k alone (very small, useless for the real world) made me wonder if it wasn't just overfit to the training set. Would it perform better training on the datasets used for the foundation models ? I doubt it.
Well I'm not saying CNNs are bad or useless at any rate.
Exactly. Most of the comparison papers are useless. This is hard stuff, only few people have the chops it takes to even attempt this. You can of course train some models and then post the numbers, that's not the hard part.
There's no 'rigorous comparison' that puts CNNs over Vits in quality and Vits unlocked more use cases easier than CNNs did. That's why they're more popular, not because it's 'bandwagon-y'.
What's the use case enabled vs running a ConvNeXt or EfficientNetV2 and using the resulting strided features as you would the resulting tokens of a ViT? I'm not saying that ViT is worse. Just saying that the scholarship around comparing them is very bad or nonexistent. You have to properly tune the hyperparam enters on both sides in a fair way, and use all the general modern training tricks also on the CNN side to make it fair.
CNNs excel in vision tasks where you have limited compute, limited memory, limited data, and want something that works super well and quick. People usually don't hook CNNs up to a transformer to get language understanding either, you have to train bespoke CNNs for specific tasks
ViTs excel where you're unbounded in compute + data and also want text understanding or have a conversation about an image
These are vibes. ViT has been shown to work fine on small data with proper hyperparam and most of what you mention is actually doable just fine with the other architecture as well.
Reality is messy. But they are the ones at least thinking the most about this, and I'd say the coin is still overall weighted in their favour than listening to an uninformed opinion.
While it's not improbable to see some theoretical realized "sovereign" individual who can wholly run and own a civilization themselves as the sole master, that is ultimately the logical progression of the empowerment of the individual through Modernism, which we have signed up for as opposed to hunter-gatherers.
You could replace AI with Advanced Education or Automation or Transhumanism really and get the same result, but I don't think there is really a "solution" beyond either staying as pets or catching up as an individual.
But at the same time, even if the rich become completely autonomous to the rest in some walled enclave, that would just resolve to the unemployed all just creating their own parallel economy again. You might loose much of the inheritance of wealth and infrastructure from the preceding civilization, but then again the preceding civilization's own norms that legacy was owned by the rich, so it's just returning back to basics.
You could also replace AI with (Proper) Education and get the same logical conclusion. The wealthy would be able to access future advanced pedagogical techniques that would equalize the advantages of both talent and hard work in a post-competence future.
But a post-competence future is ultimately a good thing for humanity. There is no inherent reason why somebody who is talented or skilled in the specific abilities that society deeems necessary should be privileged if those needs can now be automated. It just sounds more like an arbitary class deemed the "skilled" complaining about the loss of their elite status against another, more entrenched elite class relying on wealth. But as a commoner, why should I favour the meritocrats who look down on me over the wealthy who might be more magnanimous in understanding their privilege?
Oh God, what a sad day to know how to read.
"Why should I favor the people that at some point had struggles like mine and has better chances to understand and empathize with me over the people who cannot understand what the expression 'paycheck to paycheck' really entails?"
Thinking that multi-millionaires/billionaires will "understand their privilege" or that they will be "magnanimous" is beyond naivite and goes straight to stupidity. Have you read any news in the past decade‽‽
The generation of elites fostered by academic "meritocracy" have little achievements and clearly are unable to solve or even perceive modern problems. In contrast, those who were fostered in closed patronage networks navigated their countries in times of great uncertainty and built many of the institutions we rely on today.
>The thing with war is that once you have it for a certain amount of time, you create a generation of people whose kids died, wife died, neighbors and family died, you have nothing to loose anymore.
You... didn't learn history from before 1945 did you?
But despite the author's veneration of 90s and 2000s, I'd argue the works from the late 2000s and early 2010s are just as sophisticated and arguably more relevant to our modern world; Project Itoh (Genocidal Organ, Harmony), Gatchaman Crowds, Beatless, alot of their themes were prescient; Because the frontier isn't hardware, it's ontology and semantics.
reply