What a load of BS. Even if initial time saving was true because of "no bugs feature" and it is not. Any software product that serves real business needs usually constantly evolves. Most of the time is spent on planning new features and integrating those into existing architecture.
Yup. It's easy to write a to-do app. What is hard is 10 years later of features and needing to add more. While also dealing with the architecture short sightedness and pitfalls. Anyone who has worked on legacy software knows, at some point, all architectural decisions eventually become architectural pitfalls, primarily because requirements and customers change. It becomes a point where, its not really the code that's the issue, its design and customers' changing requirements.
I'm sure TSMC has a lot of influence over the Taiwanese government. However, the people of Taiwan aren't stupid in all of this. They know this is a horrible deal for them and would make sure their politicians hear about it.
Even that understates things. TSMC is Taiwan’s singular Trump card in geopolitical negotiations. Now that righteousness and moral high grounds don’t matter any more, it’s the only thing keeping Taiwan safe.
I could not give a flying fuck what it was rendered with , tech level is not impressive. But what a story and presentation. I find Flow beautiful. I, my daughter and my grandson watched it and could not take our eyes away from it.
You may not care, and you don't have to, but certainly it means a lot to the Blender Foundation, who for 23 years have been actively working on something free & open-source, and now finally it is in the big leagues.
Just as Flow's win looks even more impressive when you look at the films it competed against, who produced them, and what resources they had, Blender has been a project competing for parity and to be taken seriously while remaining totally free, and going up against systems that are either wildly expensive or not available outside the studio that made it at all.
Flow is not good because it was made with Blender, but Blender is proven to be very good and in that top echelon because Flow was made with it. For those who make or use Blender, this is big. Those folks have already believed for years/decades that Blender was great and serious, but now a lot more people outside that circle will know this, too.
Look, I am in no way trying to diminish work developers put in Blender. It is great product. I saw videos made in Blender that looks way better than Flow from the tech point of view.
My point was that the value of Flow is in its story, both written and visual and far overshadows any technical aspect. Avatar for example is totally opposite from that point of view in my opinion. Great graphics and absolutely meh, story.
And the value of Blender (or any 3D rendering software) is not only the fidelity of the rendered result, but the tools it gives artists to transform their vision into something that can be rendered by a computer.
The point is that this amazing story could be produced, as a finished movie, by a team that only needed to raise about $3M. Precisely how much of that is because they used Blender is still not clear to me, but the importance of Blender's use here is that it opens the door to equivalently great (or even better!) story telling on (relatively) low budgets ($3M is still a lot to raise, it seems to me). This story would never have been made if it needed $30M or $300M to make.
>" but the importance of Blender's use here is that it opens the door to equivalently great (or even better!) story telling on (relatively) low budgets"
I agree that what they have achieved for "only $3M" is nothing but amazing. I have no idea how much money was saved by using Blender.
> I could not give a flying fuck what it was rendered with , tech level is not impressive.
It's pretty impressive to me that something of Flow's quality could be created with free software that's avilable to anyone with an internet connection. There are a lot of highly creative people out in the world without massive amounts of money for expensive hardware/software. It's exciting for the future of animation, and I hope all the news stories talking about Flow being made with Blender will inspire more people to give it a try and see what they can do with it.
>"Competition has accelerated immensely and the final race to AGI is afoot"
And of course Sergey needs to win since poor thing need another yacht for Mondays. And of course those ungrateful workers must work their asses off to help the "needy".
This is part of what I do for living. C++ backend software running on real hardware which is currently insanely powerful. There is of course spare standby in case things go South. Works like a charm and I have yet to have a client that scratched it anywhere close to overloading server.
I understand that it can not deal with FAANG scale problems, but those are relevant only to a small subset of businesses.
>"I also agree with researchers like Yann LeCun or François Chollet that deep learning doesn't allow models to generalize properly to out-of-distribution data—and that is precisely what we need to build artificial general intelligence."
I think "generalize properly to out-of-distribution data" is too weak of criteria for general intelligence (GI). GI model should be able to get interested about some particular area, research all the known facts, derive new knowledge / create theories based upon said fact. If there is not enough of those to be conclusive: propose and conduct experiments and use the results to prove / disprove / improve theories.
And it should be doing this constantly in real time on bazillion of "ideas". Basically model our whole society. Fat chance of anything like this happening in foreseeable future.
Excluding the realtime-iness, humans do at least possess the capacity to do so.
Besides, humans are capable of rigorous logic (which I believe is the most crucial aspect of intelligence) which I don’t think an agent without a proof system can do.
Uh, if we do finally invent AGI (I am quite skeptical, LLMs feel like the chatbots of old. Invented to solve an issue, never really solving that issue, just the symptoms, and also the issues were never really understood to begin with), it will be able to do all of the above, at the same time, far better than humans ever could.
Current LLMs are a waste and quite a bit of a step back compared to older Machine Learning models IMO. I wouldn't necessarily have a huge beef with them if billions of dollars weren't being used to shove them down our throats.
LLMs actually do have usefulness, but none of the pitched stuff really does them justice.
Example: Imagine knowing you had the cure for Cancer, but instead discovered you can make way more money by declaring it to solve all of humanity, then imagine you shoved that part down everyones' throats and ignored the cancer cure part...
>"If the US is going to maintain a credible deterrent against China then something has to change. Either defense spending has to go up or we have to drastically scale back activities in other areas. And no, cheap AI drone swarms won't replace the capabilities of something like a B-21."
Assuming the US would actually need B21 capability in a war with China. Those will be probably blown up from the sky very fast. Besides I doubt wars with China and / or Russia will be limited to conventional means. Will probably escalate to nuclear very fast and then everybody is royally fucked.
Nah. Everything we know about the B-21 indicates that it's probably pretty survivable against the Chinese air defense system. Especially for stand-off strikes near the Taiwan Strait where it wouldn't have to overfly radar stations. The design was literally optimized for exactly that purpose.
Ironically the B-21 is probably safest in the air. The greatest kinetic threat is on the ground because forward air bases generally lack hardened aircraft shelters or effective missile defense. This is another reason why maintaining deterrence against China will require a major increase in defense spending or realignment of priorities.
The whole point of procuring a platform like the B-21 is to never have to use it. The strategic calculus is that just having it gives the US a range of conventional options short of global thermonuclear war, and thus forces adversaries to be more cautious.
It's hard to believe the B-21 would ever be stored at a forward missile base.
They'd likely take off from the US, fly wherever they need to go for their mission without landing, and then fly all the way back to the US. Same as the B-2.
Similarly it's unlikely a drone strike is an issue due to distance.
Believe it. Forward basing would be the only way to generate useful sortie rates in that type of conflict. Transiting even from Hickam just takes too long and requires too much tanker support. We're already short of tankers; they're wearing out and not being replaced quickly enough.
The greatest kinetic threat is from drone swarm attacks, which can be delivered to the vicinity of an airbase from a handful of anonymous-looking trucks or vans.
One or two grenade-strength shrapnel devices are enough to make a $700m dollar special project useless.
You could try a surprise drone attack from a container ship. Looks like the closest ship is about 125 miles away, that might be pushing the range a bit.
I think any war with China would be predominantly economic and cyber. I don’t think it’s viable for either party to perform a mainland strikes and I don’t think it would become nuclear, but who knows. If things escalate during the current POTUS’s term, I think it will come down to meetings and deal-makings, because that is POTUS’s passion and preferred way.
I think any war with Russia would be, well is, economic and cyber and proxy wars.
> I think it will come down to meetings and deal-makings, because that is POTUS’s passion and preferred way
Current POTUS is an egomaniacal moron. Things could very easily get out of his control, especially as he dismantles the US government and sources of influence throughout the world.
> Those will be probably blown up from the sky very fast.
Why do you say that? I'd think B21 is better (stealthier) than B2. I think there is a story about a B2 dropping a dud on NK undetected granted NK vs. China
I normally do not read blogs. Yours is different. I loved your stories. Salut to your dad. Life well lived. Also thanks for pointing to Vivian Della Chiesa. Listened few songs on Youtube - great artist.
