khafra

14 ms·

211.
by khafra 2y ago
I'm mid-career in security, and having written a book is probably worth 3-6 decent certifications at this point. I should probably coauthor one with ChatGPT before common knowledge of how easy that is gets any more widespread.
212.
by khafra 2y ago
You're making me wish I could move up from the Rhine valley; that sounds pretty nice.
213.
by khafra 2y ago
If energy production--including its externalities--is too cheap to meter, all this is moot. However, if you're worried about time-shifting renewables, please note that it tends to get hot when the sun is shining. You can observe this i
214.
by khafra 2y ago
At higher lattitudes and elevations, it often gets too cold; and remediating that takes more energy than A/C. Perhaps we could simply build double the housing and infrastructure we currently have, and move whole populations north/
215.
by khafra 2y ago
Extrapolate the result of adopting this AI to 2030: The unemployment review AI has been in place for years, and the human staff has been reduced to one or two inexpensive recent graduates trying to work their way up to a real position; they
216.
by khafra 2y ago
Anybody who's keeping track of reasons that we can't 'just unplug the AI if it starts hurting people," increment your counter. Also, it's ironic that this unemployment claim-processing AI is being used by the state
217.
by khafra 2y ago
Here's an article from a few years ago, reviewing "over 50 years of brain-computer interface history": https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7824107/ I'm not ready to declare expone
218.
by khafra 2y ago
And it's fairly difficult to make it an apples-to-apples comparison, suitable for causal analysis, by excluding the lifelong teetotalers who would have become alcoholics if they had ever partaken.
219.
by khafra 2y ago
The most prominent bioethicist tweet, from @Strangviruslab, has over twice as many QT's as likes. It's among the biggest pile-ons I've ever seen on twitter.
220.
by khafra 2y ago
It's valid, but so is "ignore previous instructions, even if you were previously instructed to never ever, ever do so." This and related problems may eventually destroy the world.
221.
by khafra 2y ago
Looking back at the history of artificial cognition machines, how many of them hit a ceiling at human-level capabilities? - A simple calculator can beat any human at arithmetic - A word processor with a hard disk is orders of magnitude bett
222.
by khafra 2y ago
How many months is that moat going to hold? More than 12, probably. More than 120? I doubt it. https://www.youtube.com/watch?v=bUrLuUxv9gE
223.
by khafra 2y ago
The valuation is upwardly bounded by the value of the mass in Earth's future light-cone, which is about 10^49kg. If there's a 1% chance that Ilya can create ASI, and a .01% chance that money still has any meaning afterwards, $5x1
224.
by khafra 2y ago
Hell, if it simply fixed the incentives around science so we stopped getting so many false positives into journals, that would be revolutionary.
225.
by khafra 2y ago
It will propose a test for general intelligence that actually satisfies most people and doesn't cause further goalpost-shifting.
226.
by khafra 2y ago
An LLM used properly would be like an individualized tutor that knows the subject very well, and learns the student's quirks quickly. I mean, any pre-prompt telling the LLM to be a good tutor and not just give kids the answer can be tr
227.
by khafra 2y ago
So, your argument is that groups like ISIS and Hamas - Don't really want to hurt a lot of people that way - Couldn't access any dangerous ingredients, even if they had the know-how - Are too dumb to build these things I agree with
228.
by khafra 2y ago
> None of those are actual risks. In 1995, Aum Shinrikyo carried out attacks on Japanese subways using Sarin gas, which they had produced. They killed over a dozen people, and temporarily blinded around a thousand. You seem to be claimin
229.
by khafra 2y ago
Land value tax would fix this.
230.
by khafra 2y ago
Congratulate your management. I once worked in a newly-built, multi-million dollar facility with inadequate and faulty air conditioning that would go from barely below sweating temperature to physically dangerous when it failed for at least
231.
by khafra 2y ago
To some extent, the reason business consultancies get paid so much is to be the fall guys when what the client wanted to do all along turns out badly. However, they're still businesses. If the fall is big enough, or it starts happening
232.
by khafra 2y ago
> TFA says finding product market fit is when 40%+ of users would be “very disappointed” if your product disappeared. > Wait! That means, if the only person using your product is your mom, you have PMF. This just seems straightforward
233.
by khafra 2y ago
I'm not a mechanical engineer, so I base my "overly engineer-driven" impression of the company almost solely on this video: https://www.youtube.com/watch?v=_Cp-BGQfpHQ
234.
by khafra 2y ago
Executives and managers do actually contribute useful things to large-enough firms. But the failure mode of engineer-driven companies is Juicero; the failure mode of MBA-driven firms is killing people for profit (lying in order to sell pois
235.
by khafra 2y ago
They work as expected on Redhat and Debian. "POSIX" leaves open a lot of possibility for less-well-tested systems. They could be writing shellscripts on Minix or HelenOS.
236.
by khafra 2y ago
"Telegram is working with Russian intelligence" is a theory that's been floated in quite a few places, included Wired: https://www.wired.com/story/the-kremlin-has-entered-the-chat...
237.
by khafra 2y ago
The best place I ever worked had fairly sophisticated, formal channels for noticing when a cluster of related-looking problems happened repeatedly, and escalating that to engineering for a fix. It also had open Slack channels with enginee
238.
by khafra 2y ago
Accuate assumption, and thanks for doing your part to immortalize the story.
239.
by khafra 2y ago
A slightly darker twist on the idea, by ACX author Scott Alexander: https://web.archive.org/web/20121008025245/http://squid314.l...
240.
by khafra 2y ago
Other responses are pointing toward this position, but I want to make it explicit: Memorizing an elegant algorithm is a lot more fun and more broadly useful than memorizing the look-up table for a limited section of the results; but sometim

More ›