theptip

51 ms·

961.
by theptip 3y ago
As a sibling noted, I think the takeaway from oxy is a lot more complex. I don’t think many people are dying from prescription meds, I think most are dying when they go to the black market to service their addiction when they cannot get pre
962.
by theptip 3y ago
The go-to example is Portugal, where drugs were decriminalized, and death rates went down. Not without challenges though, as they are now seeing issues with public drug taking and associated crime (which could be more about pandemic homeles
963.
by theptip 3y ago
Legalization & certification of quality is a good endpoint, if you can buy cheap heroin of good purity and stable dosage, no need for underground sources laced with fentanyl of variable purity. I’m here for the FDA regulating what is on
964.
by theptip 3y ago
It’s a concession to not seem weak on illegal immigration. If illegal immigrants were pardoned that would be ammo for the Republicans in election season.
965.
by theptip 3y ago
It’s true, but every step towards eventual (and IMO inevitable) legalization should be celebrated.
966.
by theptip 3y ago
No. Criminalization does more harm than benefit. For the vast majority of drug users that are not addicted, this is obvious. And for the addicts, criminalization is the last thing they need. Deterrence might prevent a few people from gettin
967.
by theptip 3y ago
I’ll bite that bullet. Legalize heroin, and nobody would want to take the more-dangerous fentanyl? But sure, legalize fentanyl and nobody would accidentally overdose. Legalize and standardize all drugs and nobody would die from fentanyl con
968.
by theptip 3y ago
The original designs had the hyperloop above ground. The later ”loop” (no hyper, drive a car onto a skate that then shunts around under the city) were underground; there might have been an early iteration of that urban design that contempla
969.
by theptip 3y ago
It’s not straightforward. What is your “X”? Duck curve shifting within the day? Or seasonal storage to account for a no-sun / no- wind black swan week like they had in Germany recently? If you are just interested in intra-day shift, th
970.
by theptip 3y ago
It’s a good question. I can see why the default behavior is the “user friendly” (but lower quality) option. But it’s confusing that there doesn’t seem to be any pro-quality dual device headsets for those willing to go through the extra hass
971.
by theptip 3y ago
The problem is, as we all discuss frequently around here, when it comes to this sort of issue social media is optimized to suppress nuance, boost controversial takes, and generate engagement through anger. So there is a very real sense in w
972.
Black Hornet: Military Micro UAV(en.wikipedia.org)
3 points by theptip 3y ago|2 comments
973.
by theptip 3y ago
> For 75 years we've been told that fusion power was only 25 years away This one always cracks me up. When I studied Physics in the early 2000s the joke was “fusion power has been 50 years away for the last 50 years”. If it’s 25 yea
974.
by theptip 3y ago
No. It specifically refers to the process of escaping the sandbox and enabling the LLM to simulate personas that the operator tried to forbid. (The sandbox is a combination of technical systems like non-AI input filtering, plus fuzzy pinky-
975.
by theptip 3y ago
> The head of SRE championed Gitlab CI. I resisted this idea because I, the relatively inexperienced manager of a nascent team, was daunted by the prospect of trying to supplant Jenkins, Github, and JIRA all at once. I think this is a re
976.
by theptip 3y ago
If your sauce is algorithmic (both current and future edge) then you cannot be OSS and profitable. Google will open source all sorts of things, but never their recommender algos. Your best bet is probably a SaaS training platform (I suspect
977.
by theptip 3y ago
Of course when testing an LLM you do not use the exact text from studies in the training set. There are an infinite number of structural permutations that preserve the methodology sufficiently to be comparable to the existing literature, wh
978.
by theptip 3y ago
This is a terrible eval. Do not update your beliefs on whether LLMs have Theory of Mind based on this paper. The eval is a weird, noisy visual task (picture of astronaut with “care packages”). Their results are hopelessly narrow. A better e
979.
by theptip 3y ago
My pride as a technologist tells me I should be able to get any Python package up and running, but man, AI dependency management is a dumpster fire right now; adding GPU driver versions into the mix seems to make everything really brittle.
980.
by theptip 3y ago
No, not at all. Just looking for the next step of many, stacking S-curves atop each other.
981.
by theptip 3y ago
My recommendation for recovering from burnout is to find the sweet spot of caring less about the work, while not finding it boring. Care 60%, not 0 or 100. “Sometimes I work hard, sometimes I just work”. It’s hard to “just work” on somethin
982.
by theptip 3y ago
IME working at a big tech company post-burnout is much more relaxing than working at a startup. There is definitely opportunity & reward for working hard and getting stuff done, but planning timelines and deadlines are so much longer. A
983.
by theptip 3y ago
The airgap option is “everybody agrees to leave lots of money on the table in the name of safety”. The alignment option is “some people invest lots of money to invent technology that everyone can use to safely make lots of money”. If you so
984.
by theptip 3y ago
Whisper (OpenAI speech-to-text) is already trained on YT content; amusingly, if you mumble incoherently, its most-probable completion for noise is “thanks for watching!”
985.
by theptip 3y ago
What? The defining trend of the last 5 or so years is the victory of the scaling hypothesis. More scale = more intelligence. GPT-4 is way smarter than 3.5, this trend is ongoing. You need more data to utilize more parameters, but folks at t
986.
by theptip 3y ago
> an attempt to regulate and control a technology that doesn't exist yet. I’m not even proposing regulation. Just trying to understand the technology. You say it doesn’t exist yet, but many AI researchers (including Sutskever and Hi
987.
by theptip 3y ago
A sufficiently intelligent system that is un-aligned can likely subvert a human operator if it wants to. But even if we ignore that, note that nobody is building their systems this way, and nor will they without extremely draconian laws req
988.
by theptip 3y ago
GP was describing acquiring power, which is completely orthogonal to being aligned with our values. (Indeed, power-seeking is usually deemed to be a bad thing.) There are certainly ecological and mimetic niches where pro-social behaviors wi
989.
by theptip 3y ago
How do you “pull the plug” on a datacenter, or on all of the cloud providers if the ASI has processes running everywhere? Given that anyone with a credit card can already achieve robust multi-region deployments, it doesn’t seem hard for an
990.
by theptip 3y ago
If AGI is achievable in 5 years, then we do need to worry now, because we currently have no idea how to solve problems like prompt hijacking or how to durably convey our values to these systems. > There is no point in worrying about it n

More ›