LESSWRONG
is fundraising!
LW

Quick Takes

"Utility" literally means usefulness, in other words instrumental value, but in decision theory and related fields like economics and AI alignment, it (as part of "utility function") is now associated with terminal/intrinsic value, almost the opposite thing (apparently through some quite convoluted history). Somehow this irony only occurred to me ~3 decades after learning about utility functions.

Showing 3 of 14 replies (Click to show all)

cubefoxnow20

I would say expected value is about how good something is in expectation. And "in expectation" means multiplied with its probability. For example, the "value" the sun shining tomorrow is 40, and its probability is 25%. So its "expected value" is 20. (This fits pretty well with Richard Jeffrey's utility theory by the way.)

2TsviBT15h

How are you distinguishing a dutch book argument from an assumption that the utility function takes some specific form (e.g. doesn't have a time index)? Cf. https://www.lesswrong.com/posts/HbkNAyAoa4gCnuzwa/wei-dai-s-shortform?commentId=SqrgPRinYbh8JCoaN (It's a genuine question, I'm not meaning to assert there isn't such a meaningful distinction; I think there is one.)

2mako yass12h

I don't know what you're asking. The answer is either trivial or mu depending on what you mean by specific form. I think if you could articulate what you're asking you wouldn't have to ask it.

soycarts's Shortform

soycarts2d-2-5

The impossible dichotomy of AI separatism

Epistemic status: offering doomer-ist big picture framing, seeking feedback.

Suppose humanity succeeds in creating AI that is more capable than the top humans across all fields, there exists a choice:

Separatism: We can consider its existence and goals as separate to humanity and seek to monitor, align, and constrain them according to our understanding of failure modes. This looks like narrow intelligence, RLHF, mechanistic interpretability. This amounts to an existential battle between human executive function and a

... (read more)

5Mitchell_Porter2d

This description is confusing, but I assume you're talking about a process in which decision-making in a human-AI hybrid ends up entirely in the AI part rather than the human part. It's logical to worry about such a thing because AI is faster than human already. However, if we actually knew what we were doing, perhaps AI superintelligence could be incorporated into an augmented human, in such a way that there is continuity of control. Wherever the executive function or the Cartesian theater is localized, maybe you can migrate it onto a faster substrate, or give it accelerated "reflexes" which mediate between human-speed conscious decision-making and faster-than-human superintelligent subsystems... But we don't know enough to do more than speculate at this point. For the big picture, your items 1 and 2 could be joined by choice 3 (don't make AI) and non-choice 4 (the AI takes over and makes the decisions). I think we're headed for 4, personally, in which case you want to solve alignment in the sense that applies to an autonomous superintelligence.

soycarts19m10

Thanks for the comment. I was imprecise with the "boundaries of reality" framing but beyond individual physical boundaries (human-AI hybrids) I'm also talking about boundaries within the fabric of social and cultural life: entertainment media, political narratives, world models. As this is influenced more by AI, I think we lose human identity.

4) to me falls under 2) as it encompasses AI free from human supervision and able permeate all aspects of life.

Eli's shortform feed

Eli Tyre13h*433

A frame to try on: If most failures of rationality are adaptively self-serving motivated reasoning, choosing to be an aspiring rationalist is basically aspiring to a kind of self-hobbling.

This is almost exactly counter to "rationality is systematized winning."

Suppose that we're living in a world where everyone is negotiating for their interests all the time. Almost everyone is engaged in motivated cognition that supports their interests in the ongoing negotiation, and that this causes them to do better for themselves on net. The rationalist is the gu... (read more)

Showing 3 of 6 replies (Click to show all)

aysja1h20

I'm not sure if I expect motivated reasoning to come out better on average, even in domains where you might naively expect it to. In part that's because self-serving strategies often involve doing things other people don't like, e.g. being deceptive, manipulative, or generally unethical, in a way that can cause long-term harm to your reputation and so long term harm to your ability to win. And I think there is significant optimization pressure on catching this kind of thing, in part for reasons similar to the ones outlined in Elephant in the Brain, i.e., t... (read more)

2Eli Tyre9h

I'll think about it. I might want to develop the idea some, first.

8Zack_M_Davis11h

Thanks, I hate it

David Matolcsi's Shortform

David Matolcsi1d5412

Someone should do the obvious experiments and replications.

Ryan Greenblatt recently posted three technical blog posts reporting on interesting experimental results. One of them demonstrated that recent LLMs can make use of filler tokens to improve their performance; another attempted to measure the time horizon of LLMs not using CoT; and the third demonstrated recent LLMs' ability to do 2-hop and 3-hop reasoning.

I think all three of these experiments led to interesting results and improved our understanding of LLM capabilities in an important safety-releva... (read more)

Showing 3 of 7 replies (Click to show all)

David Matolcsi3h20

Interesting. My guess would have been the opposite. Ryan's three posts all received around 150 karmas and were generally well-received, I think a post like this would be considered 90th percentile success for a MATS project. But admittedly, I'm not very calibrated about current MATS projects. It's also possible that Ryan has good enough intuitions to have picked two replications that are likely to yield interesting results, while a less skillfully chosen replication would be more likely to just show "yep, the phenomenon observed in the old paper is still t... (read more)

1zroe111h

I commented something similar about a month ago. Writing up a funding proposal took longer than expected but we are going to send it out in the next few days. Unless something bad happens, the fiscal sponsor will be the University of Chicago which will enable us to do some pretty cool things! If anyone has time to look at the proposal before we send it out or wants to be involved, they can send me a dm or email ([email protected]).

1loic11h

Do you have suggestions for other particularly approachable but potentially high impact replications or quick research sprints? Redwood has project proposals, but these seem higher effort than what you're suggesting and more challenging for a beginner https://blog.redwoodresearch.org/p/recent-redwood-research-project-proposals

Erich_Grunewald's Shortform

Erich_Grunewald5h30

Has anyone else noticed a thing recently (the past couple of days) where Claude is extremely reluctant to search the web, and instead is extremely keen to search past conversations or Google Drive and other nonsense like that? Even after updating my system prompt to encourage the former and discourage the latter, it still defaults to the latter. Also, instead of using the web search tool it will sometimes try and fail to search using curl and code execution. Is this just me or is anyone else experiencing similar issues?

Ian McKenzie's Shortform

Ian McKenzie20h2011

John Wentworth says:

You will never find a $100 bill on the floor of Grand Central Station at rush hour, because someone would have picked it up already.

Are you really less likely to find $100 in Grand Central Station than finding $100 anywhere else? It's true that there are many more people who could find it before you, but there are also many more people that could drop $100. If you imagine a 1D version, where everyone walks through either Grand Central Station or a quiet alley along the same line, one after the other, then it seems like you should be... (read more)

13habryka19h

Agree. Also, the one time I found a 50€ bill on the ground it was at a busy train station, so my guess is the evidence actually goes the other way around (you have many people coming through, all of whom are paying uniquely little attention to their surroundings, and so the aggregate likelihood of someone dropping a $100 bill and no one noticing before you do, is higher).

Elliott Thornley (EJT)6h10

Yeah I think the only thing that really matters is the frequency with which bills are dropped, and train stations seem like high-frequency places.

4omegastick9h

In this scenario, are you not also also paying uniquely little attention to your surroundings (and thus equally less likely to spot the bill)? It feels a little like begging the question to apply that modifier to other people in the scenario, but not yourself.

faul_sname's Shortform

faul_sname3d1316

Today, I needed to work through substantial project with a lot of drudgery (checking through an entire 1M+ LOC codebase for an http api for patterns which could cause state leakage between requests if we made a specific change to the request handling infrastructure. This involved a mix of things which are easy to do programmatically and things which require intelligent judgement, and has a fairly objective desired artifact (a list of all the places where state could leak, and a failing functional test demonstrating that leakage for each one).

I decided to d... (read more)

Showing 3 of 9 replies (Click to show all)

faul_sname6h30

the AI will often just repeat the code block in B and C so that it doesn't have to delete anything in A

Some human devs do this too. In the short term it reduces the likelihood of breaking things because something you weren't aware of relied on the old version. In the long term it makes changes harder, because now if you want to change the logic instead of changing it in one place you have to change it in n similar but usually not identical places, and if those places are different in ways that affect the implementation, now you have to try to make an in... (read more)

1Smaug12322h

I would strongly recommend asking Claude to set up a deterministic system to do this. Depending on precisely how flexible you need to be, Semgrep might be the correct tool for the job, and Claude should be able to configure that. It may be able to suggest other deterministic approaches using tools I don't know about.

2faul_sname20h

In my experience, semgrep does not play well with trying to find cross-class behavior in dynamically typed codebases with lots of dependency injection, which is why I was trying to make Claude write some code which combined static analysis (in the form of reflection or ast parsing) with runtime logic for gathering information which is hard to determine statically but easy to determine at runtime. For reference the code I ended up writing for this part was about 40 lines, it wasn't very complicated. Trying to do it in full generality purely by static analysis would be insanely complex (because php has terrible constructs like $$foo = "bar" and $foo->$bar = "baz", which this codebase doesn't use and can be trivially verified not to use, but which would be a nightmare to handle if they were used), but fortunately this wasn't what I needed. But yeah, I also expected Claude to be able to do this trivially. It is able to trivially do most tasks which feel, to me, to be about this difficult or even a bit more difficult. This task felt like it should have been easier, since it's one where there's a lot of available signal to self-correct if you make a mistake, much more so than for many of the "build and test a feature" style tasks that Claude regularly does with no drama. Which is why I thought it would be a good example for a post along the lines of "many people use LLMs to quickly add sloppy features to their codebase, increasing technical debt, but it's also possible to use them to resolve technical debt much faster than doing it by hand". And then I tried it.

Mo Putera's Shortform

Mo Putera7h90

Interesting quote on the downstream consequences of local speedup of output production by LLMs in business processes by Rafa Fernández, host of the Protocols for Business special interest group (SIG), from his essay Finding Fault Lines within the Firm:

AI is usually discussed in terms of automation or productivity. Those framings are not wrong, but they miss what makes AI adoption particularly revealing from a protocol perspective. While much of the public discussion frames AI in terms of cost-savings or new markets, our SIG has been focusing on the pressur

... (read more)

leogao's Shortform

leogao1d103

i miss the og mealsquares

3habryka19h

Nah, the new mealsquares are so much better. I love them.

leogao7h20

imo, the new mealsquares have taste and mouthfeel very similar to many brands of protein bar, whereas the old mealsquares had a unique taste and mouthfeel

4Elizabeth22h

I'm with you. My friend got the recipe by asking Romeo for it and we've made them ourselves, although it's complicated.

Zach Stein-Perlman's Shortform

Zach Stein-Perlman8d40

Concept: epistemic loadbearingness.

I write an explanation of what I believe on a topic. If the explanation is very load-bearing for my belief, that means that if someone convinces me that a parameter is wrong or points out a methodology/math error, I'll change my mind to whatever the corrected result is. In other words, my belief is very sensitive to the reasoning in the explanation. If the explanation is not load-bearing, my belief is really determined by a bunch of other stuff; the explanation might be helpful to people for showing one way I think I thin... (read more)

Showing 3 of 5 replies (Click to show all)

2Zach Stein-Perlman8d

The move I was gesturing at—which I haven't really seen other people do—is saying as opposed to the usual use of "cruxiness" which is like or

2Raemon6d

fwiw I think Lightcone staff regularly say "this is / isn't cruxy for me". (But not sure how common it is elsewhere) But, a thing I initially interpreted this post as saying that felt interesting was having more of a habit of which bits are cruxy and which are more like interesting background. (Thinking another 5 min I don't think this is that helpful a habit because when I'm writing posts I'm basically always already writing stuff that was cruxy for me)

Viliam7h20

Some things are more legible than others. If I believe something based on dozen pieces of evidence all pointing in the same direction, removing one piece of evidence wouldn't significantly change the outcome.

(Of course, removing all of them would change my mind; and even removing a few of them would make me suspicious about the remaining ones.)

So sometimes it makes sense to write things that are not cruxy.

Richard Ngo's Shortform

Richard_Ngo13d31-18

I was recently looking at the Astra Fellowship program, and found myself noticing skepticism that much value would come out of their "Strategy" category of research. In order to channel that skepticism productively, I want to talk a bit about how I think people might do valuable strategy research.

The most important thing is something like: trying to orient to the possibility of big changes in our ontologies going forward. What I mean by "strategy" is in significant part "ideas and plans that are robust to paradigm shifts".

Having said that, there's no way t... (read more)

Showing 3 of 17 replies (Click to show all)

4Richard_Ngo3d

It seems like one crux might be that I place much less value on "articulating a case" than you do. Or maybe another way of putting it is that you draw a cleaner boundary between "articulating a case" and "actually do technical work", whereas I think of them as pretty continuous when done well. (Note that this also puts me in disagreement with many rationalists. Many rationalists treat "the case for high P(doom)" as a pretty reliable set of ideas, and then "alignment research" as something we're very confused about. Whereas from my perspective, these two things are intimately related—developing the concepts and ontology required for a robust case for high P(doom) would actually get us most of the way towards solving the alignment problem.) Not quite sure how to justify my position here, but one intuition is something like "taking crucial considerations seriously". It really does seem that people's overall conclusions about what's good or bad can be flipped by things they haven't thought of. For example, I notice that many people pay lip service to the idea that the AI safety movement has significantly accelerated AI capabilities (and that AI governance has polarized the current administration against AI safety), but almost nobody is actually trying to figure out how to systematically avoid making additional mistakes like that. So how do you make your conclusions and strategies less fragile? Well, here are some domains in which it's possible to draw robust conclusions: physics, statistics, computer science, etc. Each of these fields have deep ontologies that have been tested against reality in many different ways, such that it's very hard to imagine their concepts and arguments being totally wrongheaded. Even then, you still have "crucial considerations" like relativity which totally change how you interpret fundamental concepts in those fields—but importantly, Newtonian mechanics was robust enough that even a reconceptualization of core concepts like "space" and "

Tom Davidson7h20

One other thing is that I'd have guessed that the sign uncertainty of historical work on AI safety and AI governance is much more related to the inherent chaotic nature of social and political processes rather than a particular deficiency in our concepts for understanding them.

I'm sceptical that pure strategy research could remove that side uncertainty, and I wonder if it would take something like the ability to run loads of simulations of societies like ours.

2Tom Davidson7h

Thanks for this! I do agree that the history of crucial considerations provides a good reason to favour 'deep understanding'. I also agree that you plausibly need a much deeper understanding to get to above 90% on P(doom). But I don't think you need that to get to the action-relevant thresholds, which are much lower. I'd be interested in learning more about your power grab threat models, so let me know if and when you have something you want to share. And TBC I think you're right that in many scenarios it will not be clear to other people whether the entity seeking power is ultimately humans or AIs -- my current view is that the two possibilities are distinct, and it is plausible that just one of them obtains pretty cleanly.

Eli's shortform feed

Eli Tyre16h*4-3

This post makes a seemingly-compelling case that GDP and CPI numbers are basically bullshit.

eg The numbers are based on contradictory models and BEA economists basically make arbitrary decisions of how to account for those contradictory models, including a whole slew of adjustments for factors that would otherwise be neglected. This opens up enormous room for fudging, so much so that the final numbers can swing wildly depending on which sets of (possibly a priori reasonable) assumptions and adjustments you use.

As the author puts it:

The so-calle

... (read more)

DirectedEvolution8h31

It's a poor argument top to bottom. I tried writing up a critique, but it's hard to know what to focus on because there are so many flaws. Are there any particular points that you found compelling that you would like input on?

4Unnamed13h

I haven't given that article a close read, but on a quick look through it I find it basically not at all compelling. It looks like it's the genre of 'this part of reality is surprisingly detailed, therefore be paranoid/nihilistic/cynical about it'. This genre of article is saying: there's this concept that you've been using, which you've been treating as a clean abstraction without really thinking much about where it can from, but if you look at where it comes from, there is a bunch of messy detail & judgments calls. And that much, often, is true. But it's written with an air of suspicion, or with explicit claims that therefore it's all just a bunch of made-up nonsense. Which does not follow. Numbers that involve judgment calls aren't in general fake/nonsense/bullshit. I regularly use subjective probabilities, Fermi estimates, etc., and I imagine that you do too. If you want, you can have the takeaway from this sort of article: "that's right, I've been using this concept without really understanding the messy detail behind it. Do I care enough about this to want to understand where it comes from?" If so, then you can go try to learn about it, using the processes that you usually use to learn about things. Try to find writing by economists explaining where inflation numbers and "real GDP" numbers come from, or some narrower question that you could dig into enough so that you have a . Have a conversation with Claude about it. Make a google doc where you think through what you would do if it was up to you to come up with number for "real GDP", or to do whatever tasks people do when they rely on "real GDP" numbers. Etc. This post of yours looks like it's kinda trying to do some of that, although the things that you'd want to learn about might not fit into a lesswrong comment, and I don't know if you'll find anyone who is sufficiently well-informed about real GDP calculations and willing to spend the time to leave a detailed comment for you to get a good object-le

StanislavKrym's Shortform

StanislavKrym15h*00

Steven Veld et al^[1] just released a new modification of the AI-2027 scenario as a part of MATS.

The main differences are the following:

The less tight race causes Elaris Labs to succeed in solving alignment with the help of NeuroMorph's mechinterp research. This results in the USG having an equivalent to Safer-2 and its descendants.
The AIs are professional forecasters.
Agent-4 escapes to China under the guise of being stolen, then cooperates with Deep-1, the AI created on DeepCent's compute. After Deep-1 is helped by Agent-4, Agent-4 is

... (read more)

habryka15h83

MATS just released a new modification of the AI-2027 scenario.

MATS doesn't release things. MATS is a training program! This is not some kind of official MATS release. I would phrase this differently (like saying "A MATS scholars just published")

Elizabeth's Shortform

Elizabeth2d413

Treaties Don't Exist

When I think of international regulation, I think treaty. But treaty doesn’t have a universal definition; whether an international agreement is called a treaty, agreement, or accord is a vibes-based decisions. Beyond that, there are forms of international cooperation that are definitely not treaties. It’s not obvious what hard choices or soft vibes will be best for AI policy, so I assembled a short list of options.

For a post about international treaties, this post takes a pretty US-centric approach. Sorry about that. If anyon... (read more)

Showing 3 of 4 replies (Click to show all)

2SeñorDingDong16h

Re: Soft Law. Today, I agree it is mostly wordcel bullshit. But in the 1950s and 1960s? Quite influential, especially with US + friends, even without the threat of 'bombs.' Many of the moral and legal norms on bioethics are the downstream result of soft-law promulgated in the postwar period. To get in this headspace, think of a UN resolution taking up space in the major national newspaper for several weeks, or even months. A world where single books and conferences routinely defined the future of fields and movements (e.g. Silent Spring) because of its relative media undersaturation.

Elizabeth15h30

why the change?

9Elizabeth19h

"[thing that clearly exists] doesn't exist" is established code for "is not the crisp category you think it is" (e.g. Why Fish Don't Exist). Additionally, the definition of murder might vary from place to place, but if there isn't a crisp definition for a given jurisdiction, we say that place lacks law and order. And since treaties are definitionally between multiple parties, those parties need to share a definition of anything material, in a way they don't need to share a definition of murder that occurs on their own soil.

adamzerner's Shortform

Adam Zerner8d*201

There is a concept related to scout mindset and soldier mindset (helpful outline) that I'd like to explore. Let's call it an "adversarial mindset".

From what I gather, both scout mindset and soldier mindset are about beliefs. They apply to people who are looking at the world through some belief-oriented frame. Someone who takes a soldier mindset engages in directionally motivated reasoning and asks "Can/must I believe?" whereas a scout asks "Is it true?".

On the other hand, someone who is in an adversarial mindset is looking through some sort of "combat-orie... (read more)

4Viliam18h

Yes, it seems like there is a difference between "inwards stubbornness" and "outwards stubbornness", whether people refuse to change their minds for reasons private or social. I know some people such that if you tell them they are wrong, they will double down and get angry at you... but if you meet them a few days later, they have updated their opinion. So it seems like they are willing to update, but not to admit that they did. Similarly, you tell some people a good idea, and they will tell you that it is stupid. The next day, they will come and propose the same idea as their own. I think many books on manipulation social skills recommend that the best way to change someone's mind about something is to let them believe that it was their own idea. Then again, maybe this is a smaller difference than it seems, and some people are just better at remembering what was their opinion yesterday, or better at convincing themselves that yesterday was different.

Adam Zerner16h40

Ah yeah, the phenomena you mention resonate with me and seem like evidence in favor of this idea that there is a distinction between soldier-oriented mindsets that fight against new ideas and ones that fight against something more social.

Jacobson's Shortform

Jacobson8d*00

Let's assume there is no such thing as true randomness. If this is true, and we create a superintelligent system which knows the location and properties of every particle in the universe, could we determine if we are in a simulation? (EDIT: to avoid running afoul of the impossibility of storing a complete description of the universe within the universe as @Karl Krueger pointed out, assume this includes approximations and is not exact). If we could, could we escape? If we could escape, is that still possible if there is such a thing as true randomness?

I am especially interested in answers to the final question.

Showing 3 of 6 replies (Click to show all)

Viliam19h20

A well-designed simulation is inescapable. Suppose that you are inside Conway's game of life, and you know that fact for sure. How specifically are you going to use this knowledge to escape, if all you are is a set of squares on a simulated grid, and all that ever happens in your universe is that some squares are flipped from black to white and vice versa?

To answer your first question, some kinds of pseudo-randomness are virtually indistinguishable from actual randomness, if you do not have a perfect knowledge of the entire universe. For example, in crypto... (read more)

1Jacobson2d

1. I am saying that if there is no true randomness, then the universe could be seen as a predictable program. If we are in a simulation seemingly random events are the output of a pseudorandom random number generator. If there is true randomness a superintelligent machine can't perfectly predict the future and test the limits of the universe to determine if it is simulated. 2. Perhaps 100% understanding of the universe is not needed to escape. This is an intriguing point. It does seem to me most possibilities for escape require detection.

2papetoast2d

1. The existence of true randomness eliminates some ways of detection of simulation, but not all of them. A simple example is detecting a bug in the simulation, which in theory doesn't need to depend on randomness at all. 2. It does seem to me that way too, but I think detection alone is very insufficient for escape, such that "If we could, could we escape?" isn't that meaningful of a question. You probably need to discuss with many additional assumptions to have an answer.

Charbel-Raphaël's Shortform

Charbel-Raphaël21h00

Shamelessly adapted from VDT: a solution to decision theory. I didn't want to wait for the 1st of April.

VET: A Solution to Moral Philosophy

By Claude 4.5 Opus, with prompting by Charbel Segerie

January 2026

Introduction

Moral philosophy is about how to behave ethically under conditions of uncertainty, especially if this uncertainty involves runaway trolleys, violinists attached to your kidneys, and utility monsters who experience pleasure 1000x more intensely than you.

Moral philosophy has found numerous practical applications, including generating endless Twit... (read more)

sarahconstantin's Shortform

sarahconstantin1d20

links 1/11/26: https://roamresearch.com/#/app/srcpublic/page/01-11-2026

https://www.science.org/doi/full/10.1126/sciadv.aaw7243 TRPM7 is involved in mechanosensing
- https://karger.com/cpb/article-abstract/19/1-4/1/207934/Direct-Mechano-Stress-Sensitivity-of-TRPM7-Channel
https://link.springer.com/article/10.1007/s12032-017-0987-1 TRPM7 is involved in the epithelial-mesenchymal transition in multiple tumor types
- https://journals.sagepub.com/doi/full/10.3233/CBM-190666
- https://aacrjournals.org/cancerres/article/72/16/4250/575870/TRPM7-Is-Required-for-B

... (read more)

Aprillion (Peter Hozák)'s Shortform

Aprillion1d10

My perception of time is like sampling from a continuous 2D plane of conceptual space, something akin to git railways but with a thickness to the lines, like electron orbital probability clouds that are dense around plausible points of view of what happened in the past and thin around conspiracy theories, as different linear mind-map perspectives of people standing around a sculpture, each only inferring what's on the other side, but together they can prune down non-overlapping minority reports, sticking to the consensus but never deleting (git) history.

My... (read more)

Hastings's Shortform

Hastings2d*3-9

Weaponized drones that recharge on power lines are at this point looking inevitable. if you missed the chance to freak out before everyone else about AI or covid, nows another chance.

https://www.ycombinator.com/companies/voltair

Showing 3 of 8 replies (Click to show all)

Archimedes1d40

My first thought was Amazon leveraging this for drone delivery.

2Hastings2d

I'm not talking about the US, it already has and uses this capability, along with israel, and I'm sure china has it too but they don't seem to use it.. I'm talking about russia, china, iran, pakistan, walmart, taiwan, isis, Micheal Reeves- and all able to take up the strategy of modifying other countries leadership via droning the leaders they don't like,

4β-redex2d

I think kamikaze quadcopter drones are bottlenecked on control right now, not power. One of the biggest innovations thus far, fiber-optic drones, are only necessary because the drones still need low-latency, active human control. Long range fixed wing kamikaze drones are usually autonomous, but even for those there were reports that taking remote control with FPV goggles can significantly increase accuracy and success rate. When AI is developed that can control a drone well enough, and runs on a chip that's economical to put on kamikaze drones, it's going to be a game changer. [1] Compared to that, I don't see how recharging en-route will change anything. For fiber-optic drones, landing and waiting for an ambush is already an established tactic. Sitting on a power line instead of the ground is going to make your drone much easier to notice. For fixed wing drones, Russia's Shaheds can already easily cover Ukraine, they don't need more range. 1. Game changer in the "fucking horrifying" sense of course. ↩︎