Logan Zoellner

Karma: 1,398

Logan Zoellner Dec 29, 2024, 3:47 PM
2 points
0
in reply to: hmys’s comment on: What happens next?
Is your disagreement specifically with the word “IQ” or with the broader point, that AI progress is continuing to make progress at a steady rate that implies things are going to happen soon-ish (2-4 years)?
If specifically with IQ, feel free to replace the word with “abstract units of machine intelligence” wherever appropriate.
If with “big things soon”, care to make a prediction?

Logan Zoellner Dec 27, 2024, 5:23 PM
2 points
−4
on: The Field of AI Alignment: A Postmortem, and What To Do About It
A policeman sees a drunk man searching for something under a streetlight and asks what the drunk has lost. He says he lost his keys and they both look under the streetlight together. After a few minutes the policeman asks if he is sure he lost them here, and the drunk replies, no, and that he lost them in the park. The policeman asks why he is searching here, and the drunk replies, “this is where the light is”.
I’ve always been sympathetic to the drunk in this story. If the key is in the light, there is a chance of finding it. If it is in the dark, he’s not going to find it anyway so there isn’t much point in looking there.
Given the current state of alignment research, I think it’s fair to say that we don’t know where the answer will come from. I support The Plan and I hope research continues on it. But if I had to guess, alignment will not be solved via getting a bunch of physicists thinking about agent foundations. It will be solved by someone who doesn’t know better making a discovery they “wasn’t supposed to work”.
On an interesting side here a fun story about experts repeatedly failing to make an obvious-in-hindsight discovery because they “knew better”.

Logan Zoellner Dec 13, 2024, 7:38 PM
4 points
0
in reply to: Steven Byrnes’s comment on: A shortcoming of concrete demonstrations as AGI risk advocacy
If I think AGI x-risk is >>10%, and you think AGI x-risk is 1-in-a-gazillion, then it seems self-evident to me that we should be hashing out that giant disagreement first; and discussing what if any government regulations would be appropriate in light of AGI x-risk second.
I do not think arguing about p(doom) in the abstract is a useful exercise. I would prefer the Overton Window for p(doom) look like 2-20%, Zvi thinks it should be 20-80%. But my real disagreement with Zvi is not that his P(doom) is too high, it is that he supports policies that would make things worse.
As for the outlier cases (1-in-a-gazillon or 99.5%), I simply doubt those people are amenable to rational argumentation. So, I suspect the best thing to do is to simply wait for reality to catch up to them. I doubt when there are 100M’s of humanoid robots out there on the streets, people will still be asking “but how will the AI kill us?”
(If it makes you feel any better, I have always been mildly opposed to the six month pause plan.)
That does make me feel better.

Logan Zoellner Dec 13, 2024, 6:37 PM
4 points
2
in reply to: Steven Byrnes’s comment on: A shortcoming of concrete demonstrations as AGI risk advocacy
It’s hard for me to know what’s crux-y without a specific proposal.
I tend to take a dim view of proposals that have specific numbers in them (without equally specific justifications). Examples include the six month pause, and sb 1047.
Again, you can give me an infinite number of demonstrations of “here’s people being dumb” and it won’t cause me to agree with “therefore we should also make dumb laws”
If you have an evidence-based proposal to reduce specific harms associated with “models follow goals” and “people are dumb”, then we can talk price.

Logan Zoellner Dec 13, 2024, 3:39 PM
5 points
0
on: A shortcoming of concrete demonstrations as AGI risk advocacy
“OK then! So you’re telling me: Nothing bad happened, and nothing surprising happened. So why should I change my attitude?”
I consider this an acceptable straw-man of my position.
To be clear, there are some demos that would cause me to update.
For example, I think the Solomonoff Prior is Malign to be basically a failure to do counting correctly. And so if someone demonstrated a natural example of this, I would be forced to update.
Similarly, I think the chance of a EY-style utility-maximizing agent arising from next-token-prediction are (with caveats) basically 0%. So if someone demonstrated this, it would update my priors. I am especially unconvinced of the version of this where the next-token predictor simulates a malign agent and the malign agent then hacks out of the simulation.
But no matter how many times I am shown “we told the AI to optimize a goal and it optimized the goal… we’re all doomed”, I will continue to not change my attitude.

Logan Zoellner Dec 11, 2024, 7:57 PM
6 points
−2
on: Why Isn’t Tesla Level 3?
Tesla fans will often claim that Tesla could easily do this
Tesla fan here.
Yes, Tesla can easily do the situation you’ve described (stop and go traffic on a highway in good weather with no construction). With higher reliability than human beings.
I suspect the reason Tesla is not pursuing this particular certification is because given the current rate of progress it would be out of date by the time it was authorized. There have been several significant leaps in capabilities in the last 2 years (11->12, 12->12.6, and I’ve been told 12->13). Most likely Elon (who has undeniably been over optimistic) is waiting to get FSD certified until it is at least level 4.
It’s worth noting that Tesla has significantly relaxed the requirements for FSD (from “hands on wheel” to “eyes on road”) and has done so for all circumstances, not just optimal ones.

Logan Zoellner Dec 3, 2024, 1:02 PM
4 points
0
in reply to: lewis smith’s comment on: How should TurnTrout handle his DeepMind equity situation?
Seems like he could just fake this by writing a note to his best friend that says “during the next approved stock trading window I will sell X shares of GOOG to you for Y dollars”.

Admittedly:
1. technically this is a derivative (maybe illegal?)
2. principal agent risk (he might not follow through on the note)
3. his best friend might encourage him to work harder for GOOG to succeed

But I have a hard time believing any of those would be a problem in the real world, assuming TurnTrout and his friend are reasonably virtuous about actually not wanting TurnTrout to make a profit off of GOOG.

You could come up with more complicated versions of the same thing. For example instead of his best friend, TurnTrout could gift the profit to an for-charity LLC that had AI Alignment as its mandate. This would (assuming it was set up correctly) eliminate 1. and 3.

Logan Zoellner Dec 2, 2024, 3:45 PM
2 points
0
on: How should TurnTrout handle his DeepMind equity situation?
Isn’t there just literally a financial product for this? TurnTrout could sell Puts for GOOG exactly equal to his vesting amounts/times.

Logan Zoellner Nov 30, 2024, 9:52 PM
3 points
0
in reply to: gwern’s comment on: China Hawks are Manufacturing an AI Arms Race
Einstein didn’t write a half-assed NYT op-ed about how vague ‘advances in science’ might soon lead to new weapons of war and the USA should do something about that; he wrote a secret letter hand-delivered & pitched to President Roosevelt by a trusted advisor.
Strongly agree.
What other issues might there be with this new ad hoced strategy...?
I am not a China Hawk. I do not speak for the China Hawks. I 100% concede your argument that these conversations should be taking place in a room that neither you our I are in right now.

Logan Zoellner Nov 30, 2024, 7:48 PM
4 points
0
in reply to: gwern’s comment on: China Hawks are Manufacturing an AI Arms Race
I would like to see them state things a little more clearly than commentators having to guess ‘well probably it’s supposed to work sorta like this idk?’
Meh. I want the national security establishment to act like a national security establishment. I admit it is frustratingly opaque from the outside, but that does not mean I want more transparency at the cost of it being worse. Tactical Surprise and Strategic Ambiguity are real things with real benefits.
A great example, thank you for reminding me of it as an illustration of the futility of these weak measures which are the available strategies to execute.
I think both can be true true: Stuxnet did not stop the Iranian nuclear program and if there was a “destroy all Chinese long-range weapons and High Performance Computing clusters” NATSEC would pound that button.
Is your argument that a 1-year head start on AGI is not enough to build such a button, or do you really think it wouldn’t be pressed?
It is a major, overt act of war and utter alarming shameful humiliating existential loss of national sovereignty which crosses red lines so red that no one has even had to state them—an invasion that no major power would accept lying down and would likely trigger a major backlash
The game theory implications of China waking up to finding all of their long-range military assets and GPUs have been destroyed are not what you are suggesting. A very telling current example being the current Iranian non-response to Israel’s actions against Hamas/Hezbollah.
Nukes were a hyper-exponential curve too.
While this is a clever play on words, it is not a good argument. There are good reasons to expect AGI to affect the offense-defense balance in ways that are fundamentally different from nuclear weapons.

Logan Zoellner Nov 30, 2024, 4:07 PM
7 points
3
in reply to: gwern’s comment on: China Hawks are Manufacturing an AI Arms Race
Because the USA has always looked at the cost of using that ‘robust military superiority’, which would entail the destruction of Seoul and possibly millions of deaths and the provoking of major geopolitical powers—such as a certain CCP—and decided it was not worth the candle, and blinked, and kicked the can down the road, and after about three decades of can-kicking, ran out of road.
I can’t explicitly speak for the China Hawks (not being one myself), but I believe one of the working assumptions is that AGI will allow the “league of free nations” to disarm China without the messiness of millions of deaths. Probably this is supposed to work like EY’s “nanobot swarm that melts all of the GPUs”.
I agree that the details are a bit fuzzy, but from an external perspective “we don’t publicly discuss capabilities” and “there are no adults in the room” are indistinguishable. OpenAI openly admits the plan is “we’ll as the AGI what to do”. I suspect NATSEC’s position is more like “amateurs discuss tactics, experts discuss logistics” (i.e. securing decisive advantage is more important that planning out exactly how to melt the GPUs)
To believe that the same group that pulled of Stuxnet and this lack the imagination or will to use AGI enabled weapons strikes me as naive, however.
The USA, for example, has always had ‘robust military superiority’ over many countries it desired to not get nukes, and yet, which did get nukes.
It’s also worth nothing AGI is not a zero-to-one event but rather a hyper-exponential curve. Theoretically it may be possible to always stay far-enough-ahead to have decisive advantage (unlike nukes where even a handful is enough to establish MAD).

Logan Zoellner Nov 30, 2024, 10:57 AM
3 points
0
in reply to: gwern’s comment on: China Hawks are Manufacturing an AI Arms Race
Okay, this at least helps me better understand your position. Maybe you should have opened with “China Hawks won’t do the thing they’ve explicitly and repeatedly said they are going to do”

Logan Zoellner Nov 30, 2024, 12:54 AM
2 points
−2
in reply to: gwern’s comment on: China Hawks are Manufacturing an AI Arms Race
What does winning look like? What do you do next?
This question is a perfect mirror of the brain-dead “how is AGI going to kill us?” question. I could easily make a list of 100 things you might do if you had AGI supremacy and wanted to suppress the development of AGI in China. But the whole point of AGI is that it will be smarter than me, so anything I put on the list would be redundant.

Logan Zoellner Nov 29, 2024, 1:18 PM
2 points
−4
on: AI #92: Behind the Curve
Playing the AIs definitely seems like the most challenging role
Seems like a missed opportunity not having the AIs be played bi AIs

Logan Zoellner Nov 26, 2024, 11:33 AM
0 points
0
in reply to: David Matolcsi’s comment on: “The Solomonoff Prior is Malign” is a special case of a simpler argument
yes

Logan Zoellner Nov 25, 2024, 3:30 PM
3 points
−2
on: “The Solomonoff Prior is Malign” is a special case of a simpler argument
This is a bad argument, and to understand why it is bad, you should consider why you don’t routinely have the thought “I am probably in a simulation, and since value is fragile the people running the simulation probably have values wildly different than human values so I should do something insane right now”

Logan Zoellner Nov 21, 2024, 12:17 PM
2 points
−1
in reply to: garrison’s comment on: China Hawks are Manufacturing an AI Arms Race
Chinese companies explicitly have a rule not to release things that are ahead of SOTA (I’ve seen comments of the form “trying to convince my boss this isn’t SOTA so we can release it” on github repos). So “publicly release Chinese models are always slightly behind American ones” doesn’t prove much.

Logan Zoellner Nov 10, 2024, 12:25 PM
1 point
−1
on: Could we use current AI methods to understand dolphins?
Current AI methods are basically just fancy correlations, so unless the thing you are looking for is in the dataset (or is a simple combination of things in the dataset) you won’t be able to find it.
This means “can we use AI to translate between humans and dolphins” is mostly a question of “how much data do you have?”
Suppose, for example that we had 1 billion hours of audio/video of humans/dolphins doing things. In this case, AI could almost certainly find correlations like: when dolphins pick up the seashell, they make the <<dolphin word for seashell>> sound, when humans pick up the seashell they make the <<human word for seashell>> sound. You could then do something like CLIP to find a mapping between <<human word for seashell>> and <<dolphin word for seashell>>. The magic step here is because we use the same embedding model for video in both cases, <<seashell>> is located at the same position in both our dolphin and human CLIP models.
But notice that I am already simplifying here. There is no such thing as <<human word for seashell>>. Instead, humans have many different languages. For example Papua New Guinea has over 800 languages in a land area of a mere 400k square kilometers. Because dolphins are living in what is essentially a hunter-gatherer existence, none of the pressures (trade, empire building) that cause human languages to span widespread areas exist. Most likely each pod of dolphins has at a minimum its own dialect. (one pastime I noticed when visiting the UK was that people there liked to compare how towns only a few miles apart had different words for the same things)
Dolphin lives are also much simpler than human lives, so their language is presumably also much simpler. Maybe like Eskimos have 100 words for snow, dolphins have 100 words for water. But it’s much more likely that without the need to coordinate resources for complex tasks like tool-making, dolphins simply don’t have as complex a grammar as humans do. Less complex grammar means less patterns means less for the machine learning to pick up on (machine learning loves patterns).
So, perhaps the correct analogy is: if we had a billion hours of audio/video of a particular tribe of humans and billion hours of a particular pod of dolphins we could feed it into a model like CLIP and find sounds with similar embeddings in both languages. As pointed out in other comments, it would help if the humans and dolphins were doing similar things, so for the humans you might want to pick a group that focused on underwater activities.
In reality (assuming AGI doesn’t get there first, which seems quite likely), the fastest path to human-dolphin translation will take a hybrid approach. AI will be used to identify correlations in dolphin language. For example this study that claims to have identified vowels in whale speech. Once we have a basic mapping: dolphin sounds → symbols humans can read, some very intelligent and very persistent human being will stare at those symbols, make guesses about what they mean, and then do experiments to verify those guesses. For example, humans might try replaying the sounds they think represent words/sentences to dolphins and seeing how they respond. This closely matches how new human languages are translated: a human being lives in contact with the speakers of the language for an extended period of time until they figure out what various words mean.
What would it take for an only-AI approach to replicate the path I just talked about (AI generates a dictionary of symbols that a human then uses to craft a clever experiment that uses the least amount of data possible)? Well, it would mean overcoming the data inefficiency of current machine learning algorithms. Comparing how many “input tokens” it takes to train a human child vs GPT-3, we can estimate that humans are ~1000x more data efficient than modern AI techniques.
Overcoming this barrier will likely require inference+search techniques where the AI uses a statistical model to “guess” at an answer and then checks that answer against a source of truth. One important metric to watch is the ARC prize, which intentionally has far less data than traditional machine learning techniques require. If ARC is solved, it likely means that AI-only dolphin-to-human translation is on its way (but it also likely means that AGI is immanent).
So, to answer your original question: “Could we use current AI methods to understand dolphins?” Yes, but doing so would require an unrealistically large amount of data and most likely other techniques will get there sooner.
What links here?
- Noosphere89's comment on LLMs Look Increasingly Like General Reasoners by eggsyntax (Nov 10, 2024, 3:05 PM; 2 points)

Logan Zoellner Nov 2, 2024, 11:40 PM
2 points
0
in reply to: tailcalled’s comment on: What can we learn from insecure domains?
Plausible something between 5 and 100 stories will taxonomize all the usable methods and you will develop a theory through this sort of investigation.
That sounds like something we should work on, I guess.

Logan Zoellner Nov 2, 2024, 9:49 PM
3 points
0
in reply to: Noosphere89’s comment on: What can we learn from insecure domains?
plus you are usually able to error-correct such that a first mistake isn’t fatal.”
This implies the answer is “trial and error”, but I really don’t think the whole answer is trial and error. Each of the domains I mentioned has the problem that you don’t get to redo things. If you send crypto to the wrong address it’s gone. People routinely type their credit card information into a website they’ve never visited before and get what they wanted. Global thermonuclear war didn’t happen. I strongly predict that when LLM agents come out, most people will successfully manage to use them without first falling for a string of prompt-injection attacks and learning from trial-and-error what prompts are/aren’t safe.
Humans are doing more than just trial and error, and figuring out what it is seems important.