Hopenope

Karma: 169

Hopenope 21 Jun 2026 6:34 UTC
3 points
0
in reply to: Daniel Kokotajlo’s comment on: Daniel Kokotajlo’s Shortform
Even if they are less religious, they have many incentives to hide it. it is a theocracy after all. There are some minor signs that they are becoming less religious: leaked photos of their children, and private events like weddings, are usually quite liberal.
Some of them have an apocalyptic mindset. they are not the majority but the previous supreme leader gave them a lot of power and recognition. there are some evidence that they are getting sidelined, but It is still too early to know the power dynamics and main objectives of the new leadership.

Hopenope 20 Jun 2026 16:54 UTC
4 points
1
in reply to: Gavin Runeblade’s comment on: Daniel Kokotajlo’s Shortform
Unlikely, in my opinion, for 3 main reasons:
1. Ethnic militias are not that strong. Azeris don’t have any militia, and Balochs and Kurds are far from their prime. 20 years ago they could do a lot, especially with air support, but nowadays they don’t have the required numbers or the fighting experience.
2. Ethnic lines are not so clear cut. Marrying other ethnicities is not a taboo, and people do it all the time. It is becoming even more common because of military service, universities, urbanization, and remote or contract work. The Supreme Leader, President, and Parliament Speaker have mixed ethnicity parentage.
3. Religion is dying and nationalism is taking its place. it is visible everywhere.

Hopenope 13 Jun 2026 20:39 UTC
33 points
5
in reply to: Daniel Kokotajlo’s comment on: Daniel Kokotajlo’s Shortform
As an Iranian, let me add some context:
1. Rationality and game theory dont always help when trying to understand ideological or theocratic governments.
2. Its easy to buy time by printing money. Of course, this causes inflation but inflation in Iran has been in the double digits for decades. Unless it somehow triggers an uncontrolled hyperinflation, it isn’t enough of a deterrent.
3. After the gulf War in 2003, the Islamic Republic decentralized its military. Even if the capital falls, theres a possibility that fighting could continue for years. I don’t think the military is as ideological as the Taliban, and theres a good chance morale will collapse the moment a ground invasion starts but in a decentralized system with multiple military and paramilitary organizations, its very hard to predict the outcome.
4. Israel also has elections in the spring, and they need tangible results. This makes it more likely that they will restart the war before October.
5. Iran has another strong playing card: the yemeni proxy (Houthis) could close the Strait of Bab elmandeb.
6. Mossad has been very successful with its operations in Iran. i am not sure if the Islamic Republic can build a nuke in future without them knowing or intervening.

Hopenope 13 Aug 2025 22:03 UTC
4 points
0
in reply to: Tomás B.’s comment on: Tomás B.’s Shortform
ED is not the only problem with finasteride. I saw a couple of cases of gynecomastia in medical school and stopped using finasteride after that. Minoxidil worked fine solo for 4 years, but applying it every night was annoying, and when I stopped using it, I went bald fast (Norwood 6 in 5 months!).

Hopenope 5 Aug 2025 19:36 UTC
2 points
0
in reply to: Vladimir_Nesov’s comment on: Vladimir_Nesov’s Shortform
There is a paper that shows the overreliance of in-context learning on superficial clues. it is from 2022, and the tested models are old. So, maybe newer ones are doing much better,but maybe it is not really learning, at least by some definitions.

Hopenope 19 Jul 2025 18:35 UTC
2 points
1
in reply to: Aaron Staley’s comment on: Burny’s Shortform
i think the IMO result is best of 32 and USAMO is not

Hopenope 16 Jul 2025 21:41 UTC
17 points
0
on: Hopenope’s Shortform
OpenAI is competing in the AtCoder world tour finals (heuristic division) with a new model/agent. It is a 10-hour competition with an optimization-based problem, and OpenAI’s model is currently at 2nd place.

Hopenope 16 Jul 2025 1:31 UTC
LW: 3 AF: 2
0
AF
in reply to: Daniel Kokotajlo’s comment on: Chain of Thought Monitorability: A New and Fragile Opportunity for AI Safety
Is optimizing CoT to look nice a big concern? There are other ways to show a nice CoT without optimizing for it. The frontrunners also have some incentives to not show the real CoT. Additionally, there is a good chance that people prefer a nice structured summary of CoT by a small LLM when reasonings become very long and convoluted.

Hopenope 21 May 2025 21:06 UTC
5 points
4
on: Gemini Diffusion: watch this space
What is the point of these benchmarks without knowing the training compute and data ? One of the main questions is their interpretability. Iterative refinement of these models may open new opportunities.

Hopenope 12 May 2025 7:55 UTC
6 points
0
in reply to: Garrett Baker’s comment on: Orienting Toward Wizard Power
People with a history of seizures are usually excluded from these kinds of clinical trials, so it is not an apple to apple comparison. the problem is that bupropion interacts with a lot of drugs. seizure rates are also highly dose dependent(10 times higher if taking more than 450 mg daily). Generally, if you’re not taking any interacting medications, are on the 150–300 mg slow-release version, and have no history of seizures, then the risk is low.

Hopenope 11 May 2025 8:11 UTC
14 points
0
in reply to: David Lorell’s comment on: Orienting Toward Wizard Power
As a doctor, I can tell you that even if you don’t have anxiety, it’s possible to develop some while taking bupropion/welbutrin. I used it personally and experienced the most severe anxiety I’ve ever had. It is also associated with a higher chance of seizures, and if you daydream a lot, it may make them worse. However, on the positive side, it often decreases inattention. Generally i like the drug , but it is not a first-line treatment for depression, and for good reasons.

Hopenope 25 Apr 2025 9:06 UTC
3 points
0
in reply to: qualiacat’s comment on: less-wronger-numb89′s Shortform
I lived for a while in a failing country with high unemployment. The businesses and jobs that pay well become saturated very quickly. People are less likely to spend money and often delay purchasing new stuff or maintaining their homes. Many jobs exist because we dont have time to do them ourselves, and a significant number of these jobs will just vanish. It is really hard to prepare for a high unemployment rate society.

Hopenope 26 Feb 2025 20:46 UTC
2 points
2
in reply to: Knight Lee’s comment on: Emergent Misalignment: Narrow finetuning can produce broadly misaligned LLMs
Overrefusal issues were way more common 1-2 years ago. models like gemini 1, and claude 1-2 had severe overrefusal issues.

Hopenope 23 Feb 2025 15:37 UTC
1 point
0
in reply to: MondSemmel’s comment on: LWLW’s Shortform
Your argument is actually possible, but what evidences do you have, that make it the likely outcome?

Hopenope 23 Feb 2025 11:43 UTC
2 points
1
in reply to: MondSemmel’s comment on: LWLW’s Shortform
the difficulty of alignment is still unknown. it may be totally impossible, or maybe some changes to current methods (deliberative alignment or constitutional ai) + some R&D automation can get us there.

Hopenope 10 Feb 2025 22:47 UTC
5 points
0
in reply to: Daniel Tan’s comment on: Daniel Tan’s Shortform
The recurrent paper is actually scary, but some of the stuff there are actually questionable. is 8 layers enough for a 3.5b model? qwen 0.5b has 24 layers.there is also almost no difference between 180b vs 800b model, when r=1(table 4). is this just a case of overcoming insufficient number of layers here?

Hopenope 22 Jan 2025 10:34 UTC
7 points
0
on: Hopenope’s Shortform
Is COT faithfulness already obsolete? How does it survive the concepts like latent space reasoning, or RL based manipulations(R1-zero)? Is it realistic to think that these highly competitive companies simply will not use them, and simply ignore the compute efficiency?

Hopenope 9 Jan 2025 19:29 UTC
4 points
1
on: Hopenope’s Shortform
I am not sure if longer timelines are always safer. For example, when comparing a two-year timeline to a five-year one, there are a lot of advantages to the shorter timeline. In both cases you need to outsource a lot of alignment research to AI anyway, and the amount of compute and the number of players with significant compute are lower, which reduces both the racing pressure and takeoff speed.

Hopenope 7 Jan 2025 17:31 UTC
15 points
0
on: Hopenope’s Shortform
What happened to Waluigi effect? It used to be a big issue, some people were against it, and suddenly it is pretty much forgotten. Are there any related research, or recent demos, that examine it in more detail?

Hopenope 28 Dec 2024 20:20 UTC
3 points
0
on: Hopenope’s Shortform
If you have a very short timeline, and you don’t think that alignment is solvable in such a short time, then what can you still do to reduce the chance of x-risk?