정치학자의 서재

투표 의향과 여론조사 특성을 고려한 선거 여론조사 통합지표 및 결과 예측 Likely Voter-Based Polling Aggregator & Prediction for South Korean Elections

[정치학자의 서재 @politicslibrary]

이번 21대 대선 결과를 예측해보았습니다. 수치들만 확인하지 마시고, 타래 바로 다음 트윗에 포함된 링크를 따라가시면 보다 자세한 사항들을 살펴보실 수 있습니다. 예측치 생성에 사용된 가정들 페이지는 곧 엡데이트할 예정입니다.

[정치학자의 서재 @politicslibrary]

@nabeelqu I think this thread is closer to where he actually stands: https://t.co/3EGuiJ3Jp6

[정치학자의 서재 @politicslibrary]

@nabeelqu I don't know whether you are implicitly lumping together Terry into this position but tbf he is not listed as a member of AHM (https://t.co/KkO1kJqOJu) and he often publishes guest posts on his blog that don't reflect his personal beliefs on the issue.

[정치학자의 서재 @politicslibrary]

@ArielKwiat tbf Terry is not listed as a member of AHM (https://t.co/R1GEhGzCRr) and he often publishes guest posts on his blog that don't reflect his personal positions. I think this thread is closer to where he actually stands: https://t.co/3EGuiJ3Jp6

[정치학자의 서재 @politicslibrary]

@Porphyrios64839 @robinhanson It means they're less inhibited to (ab)use sick leave for allegedly non-medical reasons given that their health care spending did not increase at the same time.

[정치학자의 서재 @politicslibrary]

@MacaesBruno @paulg I can't speak about the UI issues you're experiencing but at least in terms of search, maybe you're not using it as effectively as you could. @nabeelqu has some good tips here: https://t.co/XKTkIMEbk8

[정치학자의 서재 @politicslibrary]

@paulg @oscarmoxon Mozart?

[정치학자의 서재 @politicslibrary]

@nabeelqu @nemodi Just kill me now.

[정치학자의 서재 @politicslibrary]

@welkerlaw @tylercowen @gabriel_zucman Many people here not understanding signaling and bounded rationality lol.

[정치학자의 서재 @politicslibrary]

It is simply wild.. WILD that it is quite likely that you will get better medical advice from your LLM than your average replacement level doctor.

[Eric Topol @EricTopol]

For medical information, general AI frontier models (Google, OpenAI, Anthropic) outperformed specialized @openevidence and @UpToDate as assessed by 12 US clinicians, randomized and blinded to which model and extensive testing/benchmarks. This was not anticipated. @NatureMedicine
https://t.co/KCH1ADfQWz

[정치학자의 서재 @politicslibrary]

@tylercowen @ATabarrok Alex secretly using X as a TylerGPT query engine.

[정치학자의 서재 @politicslibrary]

@paulg Although IIRC Patrick was 19 so technically outside your scope condition, but even with that said John was 17.

[정치학자의 서재 @politicslibrary]

@paulg Aren't the Collison brothers a counterexample to this though? They both don't seem to have suffered from a lack of curiosity or narrow-mindedness even when they were full speed ahead on their startup.

[정치학자의 서재 @politicslibrary]

@JesusFerna7026 "The only way forward is to break the collective action problem: get universities to curb grade inflation (I am open to suggestions on how; caps on A’s bring their own troubles)" <-Scott Kominers recently shared a mechanism design proposal on exactly this: https://t.co/uGVthmpLWL

[정치학자의 서재 @politicslibrary]

@deanwball How much of the early system card release + announcement to withhold Mythos do you think has to do with Anthropic preempting an OpenAI major model release which would undoubtedly have capabilities in the same ballpark? That way if OAI chooses to push it out A looks responsible

[정치학자의 서재 @politicslibrary]

@JesusFerna7026 Yes, Aquinas was the absolute GOAT of steel-manning an opponent's argument before systematically dismantling it. Summa is such an underrated gem.

[rohit @krishnanrohit]

Writers who are outcome-focused, who write solely to produce an output have no problem using AI, see marketing or PRDs or strategy docs, but writers who write essays or fiction, who write carefully, what's written matters to the reader, where writing is more than a simple message to be conveyed, they hate AI.

Code just needs to work, and the level of expressivity in a line of code is limited. The audience for coders is not other coders, if so they would care a lot more about elegance, see admiration for Karpathy's work. Most coders are doing a job, coding is a means to an end, not the end itself.

A well written piece however conveys quite a lot of insight, far more than just the individual content of its words, it conveys in its affect, in the negative space, in how it uses language itself, the word choices. It's a much much richer representation of the world. It's a more complex way we connect with other minds, and in those types of writing the output isn't simply a thing to be conveyed.

[Chris Arnade 🐢🐱🚌 @Chris_arnade]

In Korea a “good seat” is so important it goes from being a pasive preference to a demand and people openly angle for them, constantly eyeing someone sitting in the best location waiting to jump into their seat it if they show any sign of leaving.

It becomes effectively a source of anxiety.

Some of that’s about smaller homes, and the intense “study culture” that makes sitting in a cafe for many hours a premium.

[정치학자의 서재 @politicslibrary]

@phokarlsson Well to be fair it was a uniquely bad morning to check your email.

[정치학자의 서재 @politicslibrary]

@stevekrouse If you can get past the surface appearances that register as red flags for most people, these actually provide a solid intro for someone dipping their toes in for the first time: https://t.co/O6mGCKFbvo

[정치학자의 서재 @politicslibrary]

@nabeelqu Yeah the revealed preference sends exactly the opposite message of what you would want if the threat to drive them out of business is to have real teeth. They’ll do just fine.