antpls

78 ms·

751.
by antpls 8y ago
A parallel meta-kialo to debate about the platform itself, like the meta-exchange website, may be a nice idea
752.
by antpls 8y ago
When you want to ensure people are only focusing on the content, and that they stay vigilant about the debate, shouldn't there be UX to make sure people don't develop habits? All debates are unique
753.
by antpls 8y ago
Maybe you could propose alternative interfaces, and let each users chose the one that best fits them.
754.
by antpls 8y ago
True, but in other languages, developers try to fill the gap with numerous libraries and languages on top of each others, with abstractions on top of abstractions. Proof assistants feel "stuck" at a low level, with high expertise
755.
by antpls 8y ago
> The cost is that the justification for the steps might become opaque to humans. If the machine can tell "this proof is correct". Does the proof really need to be reviewed by humans? As long the human can easily read the probl
756.
by antpls 8y ago
> The difference is that in a proof assistant, you cannot gloss over a detail. So many of the lines establishes well-known facts which a mathematician has internalized in their brain. Maybe an intermediate DSL that would allow some &quot
757.
by antpls 8y ago
I would like to add that the "agree"/"disagree" or "pro"/"cons" should have a random placement (left or right) and also random colors every time the page is loaded. "Cons" shouldn&
758.
by antpls 8y ago
If arguments are too formal, people won't be able to naturally share them. It would become programming in a argumentation DSL. Maybe new NLP algorithms will be able to give you a mapping from natural language to a more formal one.
759.
by antpls 8y ago
I'm kind of the cynical type guy too, but that looks like an interesting initiative to me. What makes you think it's the people? Do you have experience with previous similar tools? Also, you have to look at it with the expectation
760.
by antpls 8y ago
Is that the same proof than the ~120 lines of code of the link? If yes, proof assistants have to be improved on conciseness, imho
761.
by antpls 8y ago
That proves the direct parent is wrong about human eyeballs being more effective than detection systems.
762.
by antpls 8y ago
> Your naive faith in software developers is touching, but this is entirely speculative. Until someone actually does it we don't know how much it will cost or how long it will take. What about using those 1.4 trillions dollars to su
763.
by antpls 8y ago
We are realistically closer to autonomous AI than you imagine. Pessimism, budget planning that didn't anticipate it, ethical concerns and human employment are slowing down the development.
764.
by antpls 8y ago
https://medium.com/syncedreview/simulations-pave-the-road-fo... If you are not convinced, you can look for "waymo simulation" in your favorite search engines.
765.
by antpls 8y ago
Please, missiles dynamics are easily predictable. You can cover your plane with detection systems, not only optical, at every angles. How would an human be better at that is beyond my understanding
766.
by antpls 8y ago
If you invest a few billions dollars to the task, you will have your specialized AI with all the sensors needed to out-perform human at this task in less than a decade. There is a reason Google employees refused to work for military.
767.
by antpls 8y ago
Genuine question : how does the pilot can see the missile if the missile is chasing the plane or under the plane?
768.
by antpls 8y ago
I mean, yes we have. This is how the Go AI was trained to beat the world champion, and it also helps to develop the autonomous cars at Waymo. Give it like 5 years and you have enough compute power to simulate many more data and situations.
769.
by antpls 8y ago
I have no idea of the actual challenge, but given the missile itself is unmanned, a human would less be likely to escape if he is under pressure, stressed, or tired than an AI. Here the AI and human would have access to the same data to mak
770.
by antpls 8y ago
AIs can be trained in simulated environments to meet many more situations than any human would in its life time of practice
771.
by antpls 8y ago
My intuition is that they could already, but too long term investment plans like the F35 plan slowed developments in that direction.
772.
by antpls 8y ago
> Our plane technology is quite useless in such a situation (and considering that alpha centauri is 5 light years away and aliens likely use the same physics 10 years is quite unlikely ;) ) You don't know, anything that would preemp
773.
by antpls 8y ago
> Last time I suggested something similar, it was pointed out that unmanned aircraft are highly susceptible to jamming. Thus, the only way to make them really resilient is to make them completely operable without any human intervention.
774.
by antpls 8y ago
> People haven't even been able field autonomous cars in silicon valley. The problem of unmanned ‎maneuvers is a lot simpler in a 3D space without any obstacles than on a road with many more uncertainties and information to process.
775.
by antpls 8y ago
I was told unmanned flying in the 3D space is a lot easier to achieve than on a 2D plane with many obstacles. There is no road barriers in the air.
776.
by antpls 8y ago
> The question there is the impact. If you just risk some thing and money it lowers the barrier for going into action, as the risk of having to handing a flag to a pilot's family goes away. Let's say an hostile space alien inva
777.
by antpls 8y ago
Would a retrofitted module make the F35 able to fly unmanly and/or remotly?
778.
by antpls 8y ago
I will maybe sound harsh, but to me it looks like the Web (JavaScript?) kiddos wanted their own half-baked solution. As they only know about the Web, then they reinvent the wheel on top of the Web. Anyway, if the proposal is useful to some
779.
by antpls 8y ago
> one of the reasons snowden sent his information to journalists was to remove his own biases to be part of the process -- he wanted journalists to help go through it all, determine what was in the public interest, what could be unnecess
780.
by antpls 8y ago
> (As does eg Google who has it as the preferred real-time deliver mechanism for sitemaps.) Again, totally random fact without any single source to prove it. WebSub basically comes from the need of that one entreprise? Does it worth maki

More ›