@HostileSpectrumi
iAccount based inUnited States
About this account
- Account based in
- United States
- Connected via
- United States App Store
Account-level information from X, not a live location or the device used for a specific post.
Former intel, now academic @NDU_CIC, @TheKrulakCenter, @SIWPSColumbia @ColumbiaSIPA, @CyberStatecraft, @ElliottSchoolGW, @PAISWarwick. Apolitical, views=own
Joined March 2017
- Tweets20.6K
- Following1.5K
- Followers9.3K
- Likes36.3K
“When you put together a new mind, it’s kind of traditional for the programmer to put a Frankenstein kink in the program, just in case. Sort of a dead man’s switch”.
Michael Swanwick told us, way back in ‘88, how we would wind up controlling sentient compute.
Allowing a defector that has remained under control of hostile foreign intelligence services to have any attention in current AI policy debates is a profound error, or openly siding with one’s enemies.
Contemplating greystoke winning the small wars and losing the meta, in the conflict between AI on the cusp of superintelligence. IYKYK, in the long echoes of 1988.
The argument that any investment into commercial offensive cyber incurs lability, due to some inevitable generalized but foreseeable risk, is precisely the sort of analytically unmoored ideological assertion that discredits entirely much of the good work needed to orient and manage complex cyber weapons and services markets (that will exist regardless of what those in the salons might believe)
The problem with meat proxies is that one inherits all of the problems of the kind of employee that would uncritically relay unexamined AI outputs into the work stream, with the unreliability of someone likely not smart enough to understand the questions & issues being solved for
Open source reporting on new RAF counterspace EW capabilities raises the question, can ground to orbit employment for intelligence denial objectives serve to deter adversaries? UK defence officials comments suggest they believe so. This is not merely an RF question (show me on the OSI model where the bad thing touched you). Further argument that one can use persistent engagement to pursue deterrence.
Wargaming as selection for general officer and flag officer rank coming as a surprise to the blob, points to the same pathologies that asserted wargaming should not be a requirement of war college courses. As if the thing had not started in kreigspeil
Watching an Eastern European guitarist perform wearing a noragi jacket in the back alleys of an old city, at nearly inhuman levels of dexterity, is an unexpected fragment of cyberpunk fusion remix.
20 plus years on, low investment virtual collection activities by industry against hacktivist / underground markets is still making headlines somehow.
At $60 for a loaf of rye bread, one would assume the NYC store is really selling a custom genetically engineered ergot experience.
Toffler future shock is hitting hardest because it is now at the tipping point for those who had believed they were comfortable thinking about the future. But we have outrun the headlights of collective culture, and narrowcast speculatives have not been able to keep up.
Putting down a milestone marker. Point estimate for AI curve by June ‘27 is to expect recursive self improvement across 75% to 80% of R&D tasks. Barring war risk disruption of compute and its materiel.
It is small consolation in the vulnpocalypse that black market exploit brokers are apparently having to reinvent themselves under pressure as much as anyone else. It is less comforting to realize that they are likely moving on a faster adaptive curve.
See adult humans arguing about laundry on this app makes one think it is all an op by Big Robot to sell more domestic limited purpose clankers.
Given intense name collisions with artificial intelligence, the national security intelligence discipline probably just ought to shift to the French term “Renseignement” to distinguish what we are talking about