As I wrote, the main point of the paper was not the specific model evaluation, but the development of a benchmark which can be used to test new models.
Good benchmark development is hard work. The paper goes into the details of how it was carried out.
Now that the benchmark is available, you or anyone else could use it to evaluate the current high-end versions, and measure how the performance has changed over time.
You could also use their paper to help understand how to develop a new benchmark, perhaps to overcome some limitations in the benchmark.
That benchmark and the contents of that paper are not obsolete until there is a better benchmark and description of how to build benchmarks.
In the RC model aircraft community a common recommendation is to create safe containers to charge large lipo batteries in. People often charge their batteries in old barbecue grills, toolboxes, sandboxes, etc.
Decent 1080p quality. Not bluray level, but getting close. Definitely ahead of every other video generator.
Video production just got a lot cheaper and requires very few skills. This is basically destroying the creative video production industry (ads, product videography, youtube content of all kinds) and probably VFX industry as well.
They've been running tests for weeks under the covert name "Unicorn" and just renamed the model to Seedance a few days ago.
edit: I'm not sure why I'm being downvoted for this, except perhaps not liking the ByteDance angle.
China produces incredibly good video models and have been in the market lead for at least a year now. All of the top video models, save for Veo 3, are from China.
In fact, the only open source video models of note are all non-American (mostly Chinese, and one Israeli model).
Perhaps you should use 'sentience' or something more precise to mean 'qualia-experiencing'. The word 'consciousness' is quite overloaded and thus trips people up who haven't thought about this as extensively as you have. In particular many people take it to mean "self-awareness" (whether correctly or not), but it seems obvious to me that there are many sentient beings which lack self-awareness but still have an internal experience.
You cannot measure consciousness. You are consciousness thinking of itself as the human. Measurement is an event appearing in consciousness done by humans.
It's like persons inside GTA talking about measuring the Samsung monitor. It makes no sense cause they can never see the monitor or locate it. They appear in the monitor.
You absolutely can measure "various aspects of consciousness" -- for example, "how much of the last 24h has this consciousness been awake" seems simple. So your definition seems kind of weak, could you be more precise?
Conversely, in your definition, is consciousness the only "thing" that you would describe as not being able to measure various aspects about it? Are there any other objects or concepts which you also cannot measure various aspects about? If yes, what differentiates those things from consciousness?
He has separate motors for vertical and horizontal flights, which simplifies the design, but creates a rather bad inefficiency, the vertical motors create lots of drag during the horizontal flight.
Maybe it's not a big deal, I'm not sure. Making motors rotate would add weight for sure, thus reducing the range.
With this config, the cruise motors and prop are optimally sized for cruise - which gives non trivial gains to both eta for propulsive motor efficiency and prop efficiency.
Vs a tiltrotor/wing/body in which the cruise motor has to do double duty as lifting motors. Given it takes anywhere from ~4-7x more power to hover (depending on disc loading) than to cruise, you can see how the motors are not in an optimal throttle/rpm band in this case. Archer's CTO Munoz has actually said this publicly.
Very similar design already used by Wing. I'm guessing they did a fair bit of analysis and modeling of the cost, range, complexity, safety, etc. etc. tradeoffs before settling on what they're using currently.
Two impressive things about Wing's design are
1) load paths are designed to break the airframe in controlled ways
2) the 4 blade props have alternating shorter and longer blades for quieter aeroacoustics
Adam Savage did a video tour of their factory recently, worth a watch
Adding a tiltrotor mechanism is surely not worth the added complexity and weight, in this case. You're right though on the added weight and drag from having separate motors and props.
Tilt-rotor on all 4 motors with an extra twist: the wing shape adds to the lift in vertical mode, so you can use smaller motors, so they're more efficient even in horizontal mode.
Both statements are false, but even if they were true, if AI can do your job, and costs 1% of you, and works 100x faster, there's no reason to pay you.
I agree with you on Tailwind, but React/Vue solve the problem of creating complex webapps. If you write HTML and modify DOM manually with JS, it only works for relatively simple projects. As soon as it becomes complex, it becomes hard to track which JS code changes what DOM.
Another thing is maintability. Working with single-file components with state management systems is just a pleasure.
I think if you are looking for slaves, you should look at American history. What products of your labor do you keep in this country if you are not a business owner?
I am not a fan of all the choices the USSR made, but you are mischaracterizing a society under siege by the capitalist west.
(Though even these obsolete models did better than the best humans and domain experts).