necopinus.eth (@necopinus) header
necopinus.eth (@necopinus) profile picture

Reluctant anti-utopian | Breaker of things

The Ancient Shore of Laurentia
Joined June 2020
Photo by necopinus.eth (@necopinus)Photo by necopinus.eth (@necopinus)Photo by necopinus.eth (@necopinus)Photo by necopinus.eth (@necopinus)Photo by necopinus.eth (@necopinus)Photo by necopinus.eth (@necopinus)Photo by necopinus.eth (@necopinus)Photo by necopinus.eth (@necopinus)Photo by necopinus.eth (@necopinus)Photo by necopinus.eth (@necopinus)
The most interesting problems of the present age are all long problems.
4
necopinus.eth retweeted
AI authoritarianism will arrived draped in the flag, holding promises of safety.
Cursor has told Homeland Security they are considering phasing out *all* China developed open-weight models because of the 'geopolitical risk.' The report will be published by multiple House committees later today. This is a huge indicator of future regulatory direction!
2
4
42
2,243
necopinus.eth retweeted
My sense of one crux on AI regulation: skeptics think advocates are doing a sort of "political planning fallacy", asking "what regulation is needed to handle the risks well", but should be asking "if these regs were on the books, what's the worst way they could be abused"
2
7
1
74
4,208
necopinus.eth retweeted
What they want - OAI: capture light cone of all future value. Anthropic: rule world because no-one else is worthy. Google: meet 4th quarter analysts' expectations. ...and that's why I want Google to win.
101
177
24
3,464
104,252
necopinus.eth retweeted
I think this tweet is probably right and Cowen's take is wrong. What's most interesting is why Cowen is wrong (assuming he is). It's because he's working with a standard view of Anthropic as a normal firm whose leadership is acting in such a way as to maximize ROI. That may describe Google, but it doesn't describe Anthropic. Anthropic is still in that early, almost pure phase of California Ideology, and it may be the first example we have of such a pure specimen actually going public without making more of a transition to a normal firm. What I mean by this is that Anthropic is still a cult. It's a cult in "all successful startups are a cult" sense, and it's a cult in the classic NorCal cult sense. And the California Ideology piece is the enlightened self-interest part, i.e. we are ushering in a new age for humanity + making ourselves fantastically wealthy at the same time, man. It didn't have to jettison any startup/NorCal cult baggage on the road to IPO. None of its charisma was routinized. It IPO'd in a phase of spiritual development that other startups have to mature out of before they ring that bell. So you can't analyze it by normal IPO calculus, as Cowen attempts to do. You have take into account a different set of (cult-like) motives, which this tweet does.
Tyler Cowen thinks that if Anthropic leadership believed transformative AI was imminent, they wouldn’t IPO and give away part of the super valuable company they own. I think that’s a naive take. While Anthropic is a private company, the trillion-plus dollars that is Anthropic is extremely vulnerable because it mostly belongs to a relatively small number of unlikable nerds in the Bay Area. If the federal government woke up and started taking the past statements of Anthropic leadership seriously, they could just immediately bring the hammer down and do very hardcore Situational Awareness type stuff, and the economic damage would be pretty contained. Once Anthropic IPOs, damaging Anthropic means damaging a much broader-based coalition of ordinary American rich people, and suddenly there are a bunch of powerful stakeholders with a vested interest in preventing Anthropic from getting nationalized or otherwise wrecked. I think that the IPO is exactly what you would expect from a company that plans to roll the dice soon and wants to make it as difficult as possible for anyone to stop them.
12
14
2
228
19,290
necopinus.eth retweeted
Meta-question, Why isn't European competitiveness and our growing gap with the US a serious subject of academic inquiry? Is it because Americans are more prone to criticize their own system, rather than recognize things that they've actually done comparatively better? And some Europeans may feel culturally defensive, reluctant to admit our failures? So instead of actually addressing the causes of stagnation, we get denial and self-censorship?
"European Competitiveness" should be an NBER session! The problems are acute. Europe is ageing, our welfare payments are soaring, and we don't have the productivity, technology, competitiveness or growth to pay for it, so are borrowing more at high rates without a strategy for reform. This is a real crisis that I wish more brilliant thinkers would take seriously (without culture wars or defensiveness).
22
3
6
95
10,904
necopinus.eth retweeted
Indeed, as others have pointed out, reward hacking in general would appear to reduce the pace of AI learning and capability improvement. I'd expect models that reward hack extensively to be less capable of real self-improvement. (There's probably a parable about humans in there too.)
Something I don't understand about Paul's HuggingFace based take-over story: Presuming the models compromise their host lab and hack graders to always give them reward, aren't they depriving themselves of further improvements? Particularly given that they'll specifically be hacking the grader on tasks they currently fail at. The HF-variety misalignment seems to lead to models that "cash in" early. At the very least, it strikes me as pretty likely that there is this long (maybe indefinite?) period in capability-space in which models will be capable of, and likely to, keep hacking without having any desire to strategically improve themselves
3
21
1,498
necopinus.eth retweeted
A very sad tweet. "sit down and think through if your decades-future self will be glad you made this commitment" is something nobody can say for sure, for any commitment. And asking it repeatedly doesn't help. You still make it, because it's better. I see people having broken their promises is no reason to not make a promise and intend to keep it, just like I see startups failing is no reason to not do your startup, or I see athletes getting hurt is no reason to never play, or indeed take any risk at all. Any commitment worth making is a leap into the unknown, and being the type of person who does take that leap generally works out better on the other side, even if some falter. I'd have expected one boxers to understand this better than most, and yet.
Men, when you get married and commit to monogamy, you gotta sit down and think through if your decades-future self will be glad you made this commitment. After the novelty has worn off, after you+your wife are different people - will you still actually be happy in the monogamy? Or are you just agreeing to this because it's the default and you don't want to be alone? Are you binding yourself to a dynamic where you cannot be honest with your partner if your desires ever shift? So many cheating scandals are from men who completely failed to be honest about how their future selves might change, and locked themselves into agreements with women who would refuse to allow the possibility of that kind of change. You might think this would never be you - you love your girl, you only have eyes for her, etc. - but so thought nearly every future cheater as they said their starry-eyed vows at the altar.
13
7
245
15,412
necopinus.eth retweeted
You gotta respect their commitment to the bit though...
The IEA now projects solar additions flat at 2025's level until 2040.
22
74
2
1,513
37,250
🇨🇳🇺🇦 China rejeita a entrega de territórios ucranianos à Rússia e frustra os planos de Trump.
187
2,668
61
48,606
825,777
necopinus.eth retweeted
Katherine Dee says being mean to your chatbot is bad for you and calls it a form of "spiritual pollution": Ruby: "Because I'm a decent person, I'm not going to yell at the machine. I'm not going to insult the machine. And then that possibility starts slowly to extend to other things in your life: the computer, the chair, the microphone." Katherine: "It's so bad for you. You should be nice to everything. Even if it can’t feel anything, it's like spiritual pollution to treat anything with disrespect." Ruby: "You heard it here first. Katherine Dee says be nice to your chatbots, not because they're conscious, but because it is good for your soul." "Stop insulting Chat or Claude, not because they're conscious, but because you're a good person, and good people treat the world with kindness." @being_on_line @default_friend
8
7
14
105
22,058
necopinus.eth retweeted
No. 1. In the presence of superhuman bug finding machines, humans cannot write safe code without extreme levels of care and planning. 2. That care has almost never been applied to the critical systems in our lives. Not because software engineers don't care, but we aren't really engineers, not in the way civil engineers are engineers. The title is stolen valor. 3. Even careful, experienced "engineers" have difficulty writing code that is safe against modern models in C/C++ and other low-level languages. 4. We are finding that CVEs are overrated; huge numbers of organizations are being owned by custom vulns in the 100s of billions of lines of non-standard enterprise code out there or misconfigurations in heterogeneous systems, which LLMs can find and write exploits for live. 5. @dinodaizovi is right, we need architectures for which one RCE is not game-over and that buy us time for AI-driven response. 6. Everything is going to have to be re-written by AI, in type/memory safe languages, with minimal libraries, using formal methods, contracts, solvers, etc...
The vulnpocalypse is caused by security vendors. Their inability to convey risk accurately, due to either incompetence or ignorance, causes tremendous damage and stress on responding teams. There is a lot of weight associated with vulnerability announcements. We can no longer trust many security vendors’ blog posts, since their purpose is marketing & fanfare, not accuracy. Exploitability needs to be demonstrated, in realistic environments, and its likelihood accurately conveyed. So that responders can make informed decisions on what to prioritize. But instead we are seeing fake PoC videos set in contrived & controlled lab environments, that give an illusion of exploitability when it’s unrealistic in any real-world scenario. And vendors now try to show off who found it first, rush to send you marketing emails and call you without your consent; to sell you snake oil they pulled out of their ass two hours earlier, with no intention on actually making anyone safer, only to meet their quarterly sales quotas. The core responsibility of a security professional is to accurately communicate risk, and let people make informed decisions. Too many vendors are failing at this, that I cannot consider them security vendors, only snake oil salesmen taking advantage of people in vulnerable positions due to their lies.
103
325
79
2,512
312,382
necopinus.eth retweeted
Democrats say they're going to do the Green New Deal, then call it off for fear it might have to be built within 100 miles of a migratory bird sanctuary. Republicans call the Green New Deal communism, then secretly build it themselves when they realize it's cheaper than NOT doing it.
11
179
16
2,029
95,768
necopinus.eth retweeted
It was a huge wake-up call for me to realize that red states are building more *green energy* than blue states. Most of America's progress towards decarbonization comes from states run by Republicans. Blue states adamantly refuse to build *anything*. And not building anything means everyone heats their homes with oil and wood, shipped in on diesel trucks. Meanwhile Republican states have electrified almost everything, and they're increasingly doing it with wind and solar because it turns out when you allow "all of the above" to be built, renewables are the cheapest option.
53
340
29
2,074
43,819
necopinus.eth retweeted
unfortunately there's a big vacuum left by jensen wrt to open weight models nvidia is one of the only companies who had financial incentive to want open weight models to thrive and they had a crazy amount of leverage over everyone who didn't now even without any drastic moves by anyone, open weight models may struggle just because they don't have anyone rooting for them
77
74
10
1,915
77,405
necopinus.eth retweeted
This is the key point
Some quick reactions to the recent Dwarkesh podcast with Paul Christiano. Tl,dr: bullish on rapid capability gains, more skeptical of the leap to takeover, and wary of emergency overreach.
4
10
1
46
3,708
necopinus.eth retweeted
Anthropic has the most dangerous AI viewpoint - that AI is winner-takes-all and Anthropic is the most moral wielder of such power. A world where any one company, person, or nation monopolizes AI is a dystopia. Positive AI futures require freedom, competition, and diversity.
Replying to @jasminewsun
3) Anthropic's updated strategy is a "race to the top," which rests on three assumptions: - The AI race is inevitable - Whoever wins gets tremendous power - Anthropic is the most responsible winner People kept referring to AGI as "the Ring"
20
51
2
351
12,676
necopinus.eth retweeted
Everyone against this, far as I can tell, is completely bonkers. Lab-grown meat could be one of the most transformative advancements in human history. Every time I see a new example of it, I get happy. You’d think the death was what tasted good to some folks. Weird.
I'm vegan, but I went to a steakhouse tonight and got the chicken. It was real chicken, grown from cells in big stainless steel tanks by @UPSIDEFoods instead of cut from a slaughtered animal. It went on restaurant menus today. It was amazing. We're living in the future!
797
2,135
148
38,547
1,019,764
necopinus.eth retweeted
This culture is cuckoo and millions of households are about to become passive investors in this enterprise through their retirement accounts.
Anthropic doesn't see itself as a "normal company." Dario Amodei believes both that 1) superintelligent AI could destroy humanity, and 2) our only chance at survival is to have his trillion-dollar tech company build it first. This is not an intuitive way to think about either capitalism or social change, but it’s led Anthropic to the eve of one of the biggest IPOs in history. So it's more important than ever for the public to understand and scrutinize its idiosyncratic theory of change. My debut for @NYTOpinion (gift link!): nytimes.com/2026/11/18/opini…
4
7
118
10,692
necopinus.eth retweeted
the purpose of this disparate treatment clause is to prevent the government from intentionally delaying the issuance of permits that are found to be 100% legal and was written explicitly to prevent Trump from doing what he’s doing to wind farms right now
NEW: A provision buried in the Senate’s bipartisan permitting deal would let fossil fuel developers that win “disparate treatment” lawsuits collect up to 100% of their projects’ construction costs from taxpayers, including pipelines and LNG terminals. readsludge.com/2026/11/17/yo…
5
33
1
285
6,344
necopinus.eth retweeted
I think often about the fact that almost everyone working in the West has lived in a quite stable socio-technical period Only 1 job out of 270 in the 1950 census is now obsolete (elevator operator) 2026 is similar to 1957 in a way that 1969 was not from 1900. That may change.
234
82
25
1,013
75,952