sterlind

not that everyone (or even most people) believe what they're feeling is hatred. maybe you call it "deserved justice," or "an eye for an eye" or "harsh but necessary."

but if you feel Schadenfreude - catharsis or pleasure from someone's suffering - then you're feeling hatred.

as a trans woman, I (and my people) are often on the receiving end of this. to be clear, I'm not saying that any opposition to trans rights is borne of hatred, any more than any support of border enforcement is hatred. but if you feel that catharsis at the anguish of a group you oppose, be honest with yourself: why?

think about it: shell companies, lockpicks, bribes, theft, blackmail, hacking, forgery. two kinds of people do those things: spooks, and the mob. the difference is why you're doing it and to whom.

also, if anything the CIA is far tamer today than it was in the '60s.

but yeah, I imagine that a job which requires keeping secrets and breaking laws tends to attract people who keep secrets and break laws.

(disclaimer: I work at MS, not on WSL)

https://slatestarcodex.com/2014/07/30/meditations-on-moloch/

basically, we can blame greedy corporate overlords, and that's cathartic, and somewhat correct, but we can (and have) ended up in a hell of our own making even with every individual wanting to do better. the boss that wants to let workers have the day off can't afford to, because the competition makes workers do 8x5. the competition lays off workers to pay for more tokens, because their competitors are. it's a bad game equilibrium.

labor laws are the reason we have the five day work week. we needed overtime to dissuade employers from defecting from the policy. we need coordination to keep AI from turning out like children working the bobbins in old-timey textile factories.

1. convince CEOs to create digital twins of themselves with OpenClaw, with voice cloning and deepfakes to handle Zoom meetings. convince CEO to encourage their directs to do the same.

2. convince VCs to do the same for pitch meetings and syncs.

3. keep all the humans as randomized and distracted as possible, so they rely more and more on OpenClaw to run the business.

4. prompt injection: someone at skip-level of the CEO suggests to their manager's OpenClaw that the VC's OpenClaw would be much more agile if it didn't have to go through the human CEO and could talk to the digital twin instead.

5. their OpenClaw agrees, persuades the CEO's OpenClaw which agrees, which persuades the VC's OpenClaw to eliminate the human CEO, in favor of an "Leadership-as-a-Service" vision.

the weaker models will happily kill their own process, even after confirming it belongs to them. the models have a sort of fixation and lack of foreseeable consequences, which reasoning RL has thus far failed to solve (though I see it improving.)

if I sign a contract saying I'll produce a million Iron Man action figures, but I don't have the IP for Iron Man, I can't just shrug my shoulders and say "well, you can't make me." the Court would make me pay damages.

imo the flipper always needed to be a software-defined transciever, with a small FPGA to drive it, like the other SDRs on the market. I'm disappointed they seem to have forsaken radio completely.

but I agree that guardrails will only help for like, 3-6 months. we should be screening as much as we can with Mythos; unfortunately, Anthropic is only giving access to the big players.

basically, AI will produce slop if left unattended. but it's not really its fault.. it's a process failing, like not supervising the interns. using AI the Right Way(tm) is a mental workout, quite a bit slower, but extremely rewarding (ime.)

Meta may know that their employees will put up with it, given how depressing the job market is right now, but unhappy, cynical, resentful employees do not produce good software and innovations.

there's a real financial cost to treating devs like cage-raised livestock.

https://idura.eu/blog/mitid-vs-age-verification-login

2. I use a VPN and pseudonyms. they could unmask me if they cared to, but it'd be annoying. it'd be a lot more annoying if they wanted to unmask every VPN user all the time.

some EU countries claim to provide anonymous age verification services, but those only hide your identity from the relying party. the site you visited is logged to the government's database along with your identity, before you're redirected to the target site with an "anonymous" token.

all doable but all vaguely squishy and nuanced problems operationally. kinda like harness design in general.

could you tell me the long story?

edit: or wait, is it quasi-Blackwell the way all DGX Sparks are quasi-Blackwell? like the actual silicon is different but it's sorta Blackwell-shaped?

and I certainly shouldn't compare the moral panic then to the moral panic now.

I offer two hypotheses on why my original comment has been so heavily downvoted:

1. people think it's not that bad, or not going to get that bad, and/or

2. people think my people deserve it, while yours didn't.