@th3rdsergeevich observe - orient - decide - act (:
Собираешь пакет информации для моба. Что он видит.
Сопоставляешь с целями.
Прикидываешь варианты действий и сортируешь по ожидаемой полезности.
Обновляешь текущее поведение.
Это в общих чертах. А формально, вон дерево поведений. Библиотеки примерно подо всё есть.
@th3rdsergeevich @rf для ИИ мобов классика https://en.m.wikipedia.org/wiki/Behavior_tree_(artificial_intelligence,_robotics_and_control)
А в чём проблема с боёвкой?
@nixCraft So, how about pausing rushing AIs? That should help both camps' agendas.
@algowatching Menslifting
When you try hard sticking to Boring #Haskell and then a new GHC comes out
I didn't like that "death by fancy autocorrect" meme at first, but then I've begun to be more accepting.
The idea of humanity auto-correcting itself (out of existence) by means of war and/or gross negligence has a certain ring to it.
Another batch of humanity auto-correction risks spotted
@Xilokar @dabbling_duck@oldbytes.space this. Also, endianness.
@cstross @dakkar @jpkmensah unless your antiquated colobox is a SPARCStation old enough to avoid being an LLVM-supported platform it should be able to host a backend.
@angusm The date is off. Fancy Autocorrect will do its correction in 2032CE.
@mickdarling @erictopol According to people in nuclear security that "0" can turn into "1" very fast, just by sheer luck. The field is underfunded and understaffed and getting worse with time.
@robcornelius @AlSweigart Garbage [feedback] in - garbage results out, classic.
@omgubuntu > a classic, deb-based version (default) and, for the first time, an immutable, snap-based build for enthusiasts to experiment with.
This should be somewhere just under the title.
@emilymbender The one does not cancel the other. Having bad incentive structures doesn't automatically prevent death by sepsis or getting accidental rock in head.
Toots as he pleases.