7 ms·

IMO an example of a good use case for an LLM, which would be otherwise very hard to search for, is clarifying vague technical concepts. Just yesterday I was tr

IMO an example of a good use case for an LLM, which would be otherwise very hard to search for, is clarifying vague technical concepts. Just yesterday I was trying to remember the name of a vague concept I’d forgotten, with my overall question being: “Is there a technical term in biology for the equilibrium that occurs between plant species producing defensive toxins, and toxin resistance in the insect species that feed on those plants, whereby the plant species never has enough evolutionary pressure to increase it’s toxin load enough to kill off the insect that is adapting to it” After fruitless searching around because I didn’t have the right things to look for, putting the above in ChatGPT gave an instant reply of exactly what I was looking for: “Yes, the phenomenon you're describing is often referred to as evolutionary arms race or coevolutionary arms race.”

exe34 1y ago
that sounds like a really great example of..... searching through vector embedding. I don't think the LLM part was technically necessary.
- agubelu 1y ago
  It's all about the UI/UX. Same as in Dropbox/rsync
  - udev4096 1y ago
    No one who extensively uses dropbox uses their UI. It's mostly either the API or rclone. It's inefficient and a complete waste of time to use any UI at all
- brokencode 1y ago
  Okay, so how does the average person search through vector embedding? I would like to try this out.
  - ForceBru 1y ago
    I don't think this is _literally_ a search through vector embeddings. LLMs store embeddings of individual tokens (usually parts of words), so a result of an actual search will be top-k embeddings and the corresponding tokens, similar to the output of a Google search. You could extract the initial matrix of embeddings from some open-weights model and find tokens closest to your query. However, it's not clear why do this. OP got coherent text, so that's not search. It's _similar_, though, because attention in LLMs basically looks for most similar tokens. So to answer the question about the term, the LLM had to create a stream of tokens that's semantically closest to the given description. Well, this is somewhat like a search, but it's not exactly the same.
  - exe34 1y ago
    I didn't say the average person should use it, I was thinking the search functionality could be implemented that way and save on burning up the planet to tame Moloc.
    - brokencode 1y ago
      If it could be implemented that way and would be helpful where Google fails, then why has nobody done this? I’d love to try this out if you can point to a product that does this. You can criticize LLMs all you want, but the fact is that they provide value to people in ways that alternatives simply don’t. The energy consumption is a concern, but don’t pretend there are viable alternatives when there aren’t.
      - exe34 1y ago
        I don't have insight on what goes on inside Google - they may well be doing this at some level, but their business isn't finding stuff, it's selling ads, so getting the search right is a very low priority. The LLM people are heavily invested in ever bigger models to keep the research money flowing in, it wouldn't make sense to release a service that undercuts that. that leaves independent actors - presumably building and maintaining an up to date database is difficult, so only the big search engines do.
  - 1y ago
- jacobr1 1y ago
  And the LLM won't be necessary either for the common search case of "dc plane crash" or whatever, but increasingly that will be asked of general assistance agent, and it will dispatch an internet search of some kind, and return the result. Even if it provides 0 additional benefit over google, switching the default search location might still occur, especially if it does provide some benefits (such as summarization)
throwaway273287 1y ago
[dead]
Slow_Hand 1y ago
Same. This is one of the few uses for LLMs that I actually find useful and that I trust. They’re very helpful for helping me ask more refined questions by getting the terminology correct.
- Cthulhu_ 1y ago
  That's the truth behind finding answers; it's not about finding the answer, it's about asking the right question.
tofof 1y ago
While I do like LLMs for these tasks, unfortunately this one failed you but was a near enough miss that you couldn't see it. What you were really looking for is the Red Queen problem/hypothesis/race, named after a quote from Through the Looking Glass, with the Queen explaining to Alice: "Now, here, you see, it takes all the running you can do, to keep in the same place." In particular, the Red Queen term is specifically the equilibrium you inquired about, where relative fitness is unchanging, rather than the more general concept of an evolutionary arms race in which there can be winners and losers. The terms 'evolutionary equilibrium' and 'evolutionary steady state' are also used to capture the idea of the equilibrium, rather than just of competition. Evolutionary arms race is somewhat tautological; an arms race is the description of the selective pressure applied by other species on evolution of the species in question. (There are other, abiotic sources of selective pressures, e.g. climate change on evolutionary timescales, so while 'evolution' at least carries a broader meaning, 'arms race' adds nothing that wasn't already there.) That said, using your exact query on deepseek r1 and claude sonnet 3.7 both did include red queen in their answers, along with other related concepts like tit for tat escalation.
- margalabargala 1y ago
  This is an incorrect response. Firstly, "Evolutionary Arms Race" is not tautological, it is a specific term of art in evolutionary biology. Secondly, "evolutionary arms race" is a correct answer, it is the general case of which the Red Queen hypothesis is a special case. I do agree with you that OP described a Red Queen case, though I would hesitate to say it was because of "equilibrium"; many species in Red Queen situations have in fact gone extinct. https://en.wikipedia.org/wiki/Evolutionary_arms_race https://en.wikipedia.org/wiki/Evolutionary_arms_race https://en.wikipedia.org/wiki/Red_Queen_hypothesis https://en.wikipedia.org/wiki/Red_Queen_hypothesis
  - tofof 1y ago
    I disagree that evolutionary arms race is a specific term of art; we have many specific terms of art but 'arms race' is a broad generalization popularized by Dawkins as a pop science writer addressing a lay audience. Actual terms of art in this area would include Red Queen, the many individually termed coevolutions (antagonistic, mosaic, host-parasite, plant-herbivore, predator-prey etc), coadaptation, coextinction, the escalation hypothesis, frequency-dependent selection, reciprocal selection, asymmetric selection, the evolutionary lag, evolutionary cycling, character displacement, Fisherian runaway, evolutionary mismatch/trap, (phylogenetic) niche conservatism, fitness landscape, Grinnellian vs Eltonian niches, the competitive exclusion principle, and on and on. All of these actual terms of art fit under the broad, general umbrella of an 'arms race' with other species, which is really nothing more than a restatement of Spencer's unfortunate phrase. The latter is so widely 'known' that it is to the point that I and many of my peers try not to utter it, in an effort to reduce the work refuting the same tired misunderstandings that arise from that verbiage. At any rate, almost NONE of these actual terms of art are about the sort of equilibrium that was the exact heart of the OP's query to the LLM, and thus nearly none of the broader umbrella 'arms race' is about why the plant doesn't have the evolutionary pressure to actually drive the parasite extinct. An arms race doesn't have to be in equilibrium. Armor vs weapons were in an arms race and indeed at equilibrium for millenia, but then bullets come along and armor goes exinct almost overnight and doesn't reappear for 5 centuries. Bullets win the arms race. Arms races have nothing to do, inherently, with equilibrium. You seem to have misunderstood the nature of the equilibrium in a Red Queen scenario, which is the fundamental effect that the hypothesis is directly named for. That species that are in Red Queen relationships can go extinct is in no way a counterargument to the idea that two (or more) species tend to coevolve in such a way that the relative fitness of each (and of the system as a whole) stays constant. See, for example, the end of the first paragraph on the origin of Van Valen's term at your own wiki link. Evolutionary steady-state is a synonymous term without the baggage of the literary reference and also avoids the incorrect connotation suggested by arms race that leads people to forget the abiotic factors that are often a dominant mechanism in extinctions as the realized niche vs the fundamental niche differ. Instead, Van Valen was specifically proposing the Red Queen hypothesis as an explanation of why extinction appears to be a half-life, i.e. of a constant probability, rather than a rate that depends on the lifetime of the taxa. This mechanism has good explanatory power for the strong and consistent evidence that speciation rate (usually considered as the log of the number of genera, depending on definition, see Stanley's Rule) has a direct and linear relation with the extinction rate. If Red Queened species didn't go exinct, Van Valen wouldn't have needed to coin the term to explain this correlation. Or were you deliberately invoking Cunningham's Law?
RyanOD 1y ago
Agreed. I'm increasingly using ChatGPT to research topics. In that way, I can refine my question, drill down, ask for alternatives, supply my own supplementary information, etc. I think of AI as an intelligent search engine / assistant and, outside of simple questions with one very specific answer, it just crushes search engines.
- schrodinger 1y ago
  it's especially great in voice mode. I love to take a long walk or jog with earbuds in and use it to learn a new topic. I don't well through indoctrination, I learn much better by asking it to give a very big picture overview, start to build an intuitive understanding, and then start asking pointed questions to fill in the blanks and iterate towards a deeper understanding. I find ChatGPT's voice mode to be very engaging in that use case—far more so than Gemini's actually—and I've used it to learn a bunch of new technologies. It's like having a never-bored professor friend you can call and ask endless questions and keep getting answers without annoying it!
  - jimmySixDOF 1y ago
    Yeah, I find setting it in voice mode and then directing it to a specific site that I'm interested in discussing is super useful, so I start off in ChatGPT*, then I switch windows to what I'm interested in looking at. For instance, I want to go over a GitHub repo. I can ask it to verbally ask it to go there and then chat with it as I read through the README file, and it provides great a soundboarding experience. No pun intended. Heck, I'm even dictating this response through Wispr Flow, which I have found to be more useful than I had anticipated. * Gemini lets you do this by actually streaming your screen through and verbally chatting about whatever is on screen. While interesting, I find the responses to be a little less thorough. YMMV.
- Shorel 1y ago
  I don't see it as crushing the search engines. I use the LLMs to find the right search terms, and that combination makes search engines much more useful. LLM by themselves give me very superficial explanations that don't answer what I want, but they are also a great starting point that will eventually guide me to the answers.
throwaway422432 1y ago
Synthesizing what would be multiple searches into one prompt is where they can be really useful. For example: Take the ingredient list of a cosmetic or other product that could be 30-40 different molecules and ask ChatGPT to list out what each of them is and if any have potential issues. You can then verify what it returns via search.
paul7986 1y ago
Present day... Google 55% as GPT is not a local search engine GPT 45% but use it for more intelligent learning/conversations/knowledgebase. If I had a GPT phone ... sorta like H.E.R. the movie I would rarely leave my phone's lockscreen. My AI device / super AI human friend would do everything for me including get me to the best lighting to take the best selfies...
Cthulhu_ 1y ago
I just want to say that this thread / the responses to your question are better than either search engines or LLMs can ever come up with, and shows the truth of Cunningham's Law: "The best way to get the right answer on the internet is not to ask a question; it's to post the wrong answer" Or updated for the LLM age, "the best way to get the right answer from an LLM is not to ask it a question and use its answer; it's to post its response on a site of well-educated and easily nerdsniped people"
- inanutshellus 1y ago
  The responses to GP's comment are surprisingly entertaining to read. I'm entirely engaged in this nerdy semantics war on a topic I know nothing about and loving it. Time to pop some popcorn and hit refresh.

exe34 1y ago

that sounds like a really great example of..... searching through vector embedding. I don't think the LLM part was technically necessary.

agubelu 1y ago
It's all about the UI/UX. Same as in Dropbox/rsync
- udev4096 1y ago
  No one who extensively uses dropbox uses their UI. It's mostly either the API or rclone. It's inefficient and a complete waste of time to use any UI at all
brokencode 1y ago
Okay, so how does the average person search through vector embedding? I would like to try this out.
- ForceBru 1y ago
  I don't think this is _literally_ a search through vector embeddings. LLMs store embeddings of individual tokens (usually parts of words), so a result of an actual search will be top-k embeddings and the corresponding tokens, similar to the output of a Google search. You could extract the initial matrix of embeddings from some open-weights model and find tokens closest to your query. However, it's not clear why do this. OP got coherent text, so that's not search. It's _similar_, though, because attention in LLMs basically looks for most similar tokens. So to answer the question about the term, the LLM had to create a stream of tokens that's semantically closest to the given description. Well, this is somewhat like a search, but it's not exactly the same.
- exe34 1y ago
  I didn't say the average person should use it, I was thinking the search functionality could be implemented that way and save on burning up the planet to tame Moloc.
  - brokencode 1y ago
    If it could be implemented that way and would be helpful where Google fails, then why has nobody done this? I’d love to try this out if you can point to a product that does this. You can criticize LLMs all you want, but the fact is that they provide value to people in ways that alternatives simply don’t. The energy consumption is a concern, but don’t pretend there are viable alternatives when there aren’t.
    - exe34 1y ago
      I don't have insight on what goes on inside Google - they may well be doing this at some level, but their business isn't finding stuff, it's selling ads, so getting the search right is a very low priority. The LLM people are heavily invested in ever bigger models to keep the research money flowing in, it wouldn't make sense to release a service that undercuts that. that leaves independent actors - presumably building and maintaining an up to date database is difficult, so only the big search engines do.
- 1y ago
jacobr1 1y ago
And the LLM won't be necessary either for the common search case of "dc plane crash" or whatever, but increasingly that will be asked of general assistance agent, and it will dispatch an internet search of some kind, and return the result. Even if it provides 0 additional benefit over google, switching the default search location might still occur, especially if it does provide some benefits (such as summarization)

throwaway273287 1y ago

[dead]

Slow_Hand 1y ago

Same. This is one of the few uses for LLMs that I actually find useful and that I trust. They’re very helpful for helping me ask more refined questions by getting the terminology correct.

Cthulhu_ 1y ago
That's the truth behind finding answers; it's not about finding the answer, it's about asking the right question.

tofof 1y ago

While I do like LLMs for these tasks, unfortunately this one failed you but was a near enough miss that you couldn't see it. What you were really looking for is the Red Queen problem/hypothesis/race, named after a quote from Through the Looking Glass, with the Queen explaining to Alice: "Now, here, you see, it takes all the running you can do, to keep in the same place." In particular, the Red Queen term is specifically the equilibrium you inquired about, where relative fitness is unchanging, rather than the more general concept of an evolutionary arms race in which there can be winners and losers. The terms 'evolutionary equilibrium' and 'evolutionary steady state' are also used to capture the idea of the equilibrium, rather than just of competition. Evolutionary arms race is somewhat tautological; an arms race is the description of the selective pressure applied by other species on evolution of the species in question. (There are other, abiotic sources of selective pressures, e.g. climate change on evolutionary timescales, so while 'evolution' at least carries a broader meaning, 'arms race' adds nothing that wasn't already there.) That said, using your exact query on deepseek r1 and claude sonnet 3.7 both did include red queen in their answers, along with other related concepts like tit for tat escalation.

margalabargala 1y ago
This is an incorrect response. Firstly, "Evolutionary Arms Race" is not tautological, it is a specific term of art in evolutionary biology. Secondly, "evolutionary arms race" is a correct answer, it is the general case of which the Red Queen hypothesis is a special case. I do agree with you that OP described a Red Queen case, though I would hesitate to say it was because of "equilibrium"; many species in Red Queen situations have in fact gone extinct. https://en.wikipedia.org/wiki/Evolutionary_arms_race https://en.wikipedia.org/wiki/Evolutionary_arms_race https://en.wikipedia.org/wiki/Red_Queen_hypothesis https://en.wikipedia.org/wiki/Red_Queen_hypothesis
- tofof 1y ago
  I disagree that evolutionary arms race is a specific term of art; we have many specific terms of art but 'arms race' is a broad generalization popularized by Dawkins as a pop science writer addressing a lay audience. Actual terms of art in this area would include Red Queen, the many individually termed coevolutions (antagonistic, mosaic, host-parasite, plant-herbivore, predator-prey etc), coadaptation, coextinction, the escalation hypothesis, frequency-dependent selection, reciprocal selection, asymmetric selection, the evolutionary lag, evolutionary cycling, character displacement, Fisherian runaway, evolutionary mismatch/trap, (phylogenetic) niche conservatism, fitness landscape, Grinnellian vs Eltonian niches, the competitive exclusion principle, and on and on. All of these actual terms of art fit under the broad, general umbrella of an 'arms race' with other species, which is really nothing more than a restatement of Spencer's unfortunate phrase. The latter is so widely 'known' that it is to the point that I and many of my peers try not to utter it, in an effort to reduce the work refuting the same tired misunderstandings that arise from that verbiage. At any rate, almost NONE of these actual terms of art are about the sort of equilibrium that was the exact heart of the OP's query to the LLM, and thus nearly none of the broader umbrella 'arms race' is about why the plant doesn't have the evolutionary pressure to actually drive the parasite extinct. An arms race doesn't have to be in equilibrium. Armor vs weapons were in an arms race and indeed at equilibrium for millenia, but then bullets come along and armor goes exinct almost overnight and doesn't reappear for 5 centuries. Bullets win the arms race. Arms races have nothing to do, inherently, with equilibrium. You seem to have misunderstood the nature of the equilibrium in a Red Queen scenario, which is the fundamental effect that the hypothesis is directly named for. That species that are in Red Queen relationships can go extinct is in no way a counterargument to the idea that two (or more) species tend to coevolve in such a way that the relative fitness of each (and of the system as a whole) stays constant. See, for example, the end of the first paragraph on the origin of Van Valen's term at your own wiki link. Evolutionary steady-state is a synonymous term without the baggage of the literary reference and also avoids the incorrect connotation suggested by arms race that leads people to forget the abiotic factors that are often a dominant mechanism in extinctions as the realized niche vs the fundamental niche differ. Instead, Van Valen was specifically proposing the Red Queen hypothesis as an explanation of why extinction appears to be a half-life, i.e. of a constant probability, rather than a rate that depends on the lifetime of the taxa. This mechanism has good explanatory power for the strong and consistent evidence that speciation rate (usually considered as the log of the number of genera, depending on definition, see Stanley's Rule) has a direct and linear relation with the extinction rate. If Red Queened species didn't go exinct, Van Valen wouldn't have needed to coin the term to explain this correlation. Or were you deliberately invoking Cunningham's Law?

RyanOD 1y ago

Agreed. I'm increasingly using ChatGPT to research topics. In that way, I can refine my question, drill down, ask for alternatives, supply my own supplementary information, etc. I think of AI as an intelligent search engine / assistant and, outside of simple questions with one very specific answer, it just crushes search engines.

schrodinger 1y ago
it's especially great in voice mode. I love to take a long walk or jog with earbuds in and use it to learn a new topic. I don't well through indoctrination, I learn much better by asking it to give a very big picture overview, start to build an intuitive understanding, and then start asking pointed questions to fill in the blanks and iterate towards a deeper understanding. I find ChatGPT's voice mode to be very engaging in that use case—far more so than Gemini's actually—and I've used it to learn a bunch of new technologies. It's like having a never-bored professor friend you can call and ask endless questions and keep getting answers without annoying it!
- jimmySixDOF 1y ago
  Yeah, I find setting it in voice mode and then directing it to a specific site that I'm interested in discussing is super useful, so I start off in ChatGPT*, then I switch windows to what I'm interested in looking at. For instance, I want to go over a GitHub repo. I can ask it to verbally ask it to go there and then chat with it as I read through the README file, and it provides great a soundboarding experience. No pun intended. Heck, I'm even dictating this response through Wispr Flow, which I have found to be more useful than I had anticipated. * Gemini lets you do this by actually streaming your screen through and verbally chatting about whatever is on screen. While interesting, I find the responses to be a little less thorough. YMMV.
Shorel 1y ago
I don't see it as crushing the search engines. I use the LLMs to find the right search terms, and that combination makes search engines much more useful. LLM by themselves give me very superficial explanations that don't answer what I want, but they are also a great starting point that will eventually guide me to the answers.

throwaway422432 1y ago

Synthesizing what would be multiple searches into one prompt is where they can be really useful. For example: Take the ingredient list of a cosmetic or other product that could be 30-40 different molecules and ask ChatGPT to list out what each of them is and if any have potential issues. You can then verify what it returns via search.

paul7986 1y ago

Present day... Google 55% as GPT is not a local search engine GPT 45% but use it for more intelligent learning/conversations/knowledgebase. If I had a GPT phone ... sorta like H.E.R. the movie I would rarely leave my phone's lockscreen. My AI device / super AI human friend would do everything for me including get me to the best lighting to take the best selfies...

Cthulhu_ 1y ago

I just want to say that this thread / the responses to your question are better than either search engines or LLMs can ever come up with, and shows the truth of Cunningham's Law: "The best way to get the right answer on the internet is not to ask a question; it's to post the wrong answer" Or updated for the LLM age, "the best way to get the right answer from an LLM is not to ask it a question and use its answer; it's to post its response on a site of well-educated and easily nerdsniped people"

inanutshellus 1y ago
The responses to GP's comment are surprisingly entertaining to read. I'm entirely engaged in this nerdy semantics war on a topic I know nothing about and loving it. Time to pop some popcorn and hit refresh.