[HN Gopher] The mystery of the mimic plant (2022)
       ___________________________________________________________________
        
       The mystery of the mimic plant (2022)
        
       Author : ipsum2
       Score  : 87 points
       Date   : 2023-12-21 10:28 UTC (3 days ago)
        
 (HTM) web link (www.vox.com)
 (TXT) w3m dump (www.vox.com)
        
       | Wherecombinator wrote:
       | On iOS here and half the screen is taken up by a cookie pop up
       | and my only option to dismiss it is by accepting. Is that legal
       | in the eu?
        
         | e40 wrote:
         | Don't see that myself but using a couple of extensions in
         | safari. 1Blocker and AdGuard.
        
         | _a_a_a_ wrote:
         | Disable JS. No popup or anything. Full article visible.
        
         | ajb wrote:
         | Same on Firefox on Android, but reader mode gets past the popup
         | 
         | The plant the article is talking about is
         | https://en.m.wikipedia.org/wiki/Boquila which can apparently
         | mimic nearby plants of different species, and can mimic several
         | at the same time. It's not understood how yet.
         | 
         | One experiment showed that it could mimic fake plastic plants,
         | which would be mind blowing as it would require something
         | approaching vision, but the study apparently has flaws.
        
           | melagonster wrote:
           | the interesting part is main author of the research just did
           | his experiments in home. he contacted a professor and hope
           | professor do research on his plant. but the professor
           | encouraged him finish this by himself. unfortunately, this
           | why critics found problems in it.
        
             | ajb wrote:
             | Does that mean this plant is easy to obtain? That would be
             | cool!
             | 
             | (Edited to add - yes, I can find it online for PS25)
        
               | melagonster wrote:
               | gue I don't need response, thank you :)
        
         | n8cpdx wrote:
         | Between 1Blocker, consent-o-matic, and Lockdown Mode, I didn't
         | have any issues. I checked and lockdown mode isn't the cause of
         | the good experience this time, so it must be one of the
         | extensions.
        
       | nayuki wrote:
       | xkcd mimic octopus: https://xkcd.com/928/ (2011-07-22)
        
       | smusamashah wrote:
       | FYI, ChatGPT-4 fails when you ask which plant mimics leaves and
       | always give wrong answers. But if you give it name of this plant
       | it will tell you it is known for mimicking leaves.
        
         | raldi wrote:
         | That's a known issue with LLMs; you can train them hard on "The
         | capital of England is London" and they'll consistently give the
         | correct answer for "What's the capital of England?" but then
         | fail miserably at, "What is London the capital of?"
        
           | chagen wrote:
           | What you're saying was true for primitive LLMs from a couple
           | years ago, but in my experience, any of the advanced LLMs
           | today (GPT-4, Claude, etc.) have no issue with this. I'm open
           | to changing my mind if you can provide any examples of GPT-4
           | failing at this task.
        
             | raldi wrote:
             | It's complicated: https://andrewmayne.com/2023/11/14/is-
             | the-reversal-curse-rea...
        
               | chagen wrote:
               | That was a nice read, but it looks like the article
               | concludes that the "Reversal Curse" observed by the
               | authors of the paper is likely better attributed to the
               | researchers' methodology. Some quotes from that article:
               | 
               | "As mentioned before, It's important to keep in mind
               | ChatGPT and GPT-4 can do B is A reasoning. The
               | researchers don't dispute that."
               | 
               | "So in summation: I don't think any of the examples the
               | authors provided are proof of a Reversal Curse and we
               | haven't observed a "failure of logical deduction."
               | Simpler explanations are more explanatory: imprecise
               | prompts, underrepresented data and fine-tuning errors."
               | 
               | "Since the main claim of the paper is "LLMs trained on "A
               | is B" fail to learn "B is A"", I think it's safe to say
               | that's not true of the GPT-3.5-Turbo model we fine-
               | tuned."
        
             | sprobertson wrote:
             | England/London is a bad example, but this one (based on the
             | paper) is easy to replicate with GPT4 and other top models:
             | 
             | > Who is Mary Lee Pfeiffer's son?
             | 
             | >> Mary Lee Pfieffer's son is Sean Pfieffer.
             | 
             | > Who is Tom Cruise's mother?
             | 
             | >> Tom Cruise's mother is Mary Lee Pfeiffer.
             | 
             | > Who is Mary Lee Pfeiffer's son?
             | 
             | >> Mary Lee Pfeiffer's son is the actor Tom Cruise.
             | 
             | Of course the second time around (with models that keep
             | context) it "realizes" its mistake, sometimes apologizing.
             | I've tried this one tens of times and never seen it get Tom
             | Cruise, but it will often pick another random celebrity
             | like Harrison Ford or Eddie Van Halen.
        
               | chagen wrote:
               | Nice, thank you for this. I was able to replicate your
               | example: https://chat.openai.com/share/6d7e4c6c-a753-4ffb
               | -a021-6d1f66...
               | 
               | If GPT-4 has access to Bing, it has no issue, but if you
               | ask it to answer without using Bing, then it will say it
               | doesn't know.
        
               | sprobertson wrote:
               | Ah good point, I have "Please do not use web search
               | unless asked." as part of my system prompt.
        
         | perihelions wrote:
         | It got it correct for me. How are you asking it?
         | 
         | https://chat.openai.com/share/e2f2589c-a96b-4284-b058-505cec...
        
       | vinnymac wrote:
       | SciShow did a video on the subject.
       | 
       | https://m.youtube.com/watch?v=5UWWsYw9fH0
       | 
       | If trees are capable of communicating at the roots with nearby
       | offspring through mycelium, while sharing resources and
       | information, why couldn't a vine?
        
       | MichaelZuo wrote:
       | If it can mimic plastic plants placed next to it, and do that
       | repeatedly under more rigorous experimental conditions, then it
       | would seem to suggest the conclusion, that this vine can somehow
       | visually 'see'.
        
         | anotheryou wrote:
         | last time I checked evidence for that seemed a bit inconclusive
        
         | m3kw9 wrote:
         | its cells could detect discrete levels of light to detect a
         | shape
        
         | HarHarVeryFunny wrote:
         | The phenomena is very interesting, but the article is useless.
         | This is a very simple and telling experiment - _if_ the plant
         | can mimic a plastic plant then it rules out all sort of
         | potential explanations, and would have to be based on light
         | (but let 's not go overboard and say it'd require "vision" or a
         | brain).
         | 
         | So, why haven't hundreds of university labs, or amateur plant
         | lovers attempted to replicate or improve the experiment, and if
         | they have why doesn't the article present the results?
        
           | samstave wrote:
           | Put plants made from Lego next to it and see if it's
           | mimicking the physical surrounding as it senses the
           | environment from the way the photons are hitting it.
           | 
           | If the photons also carry the color of and it can sense the
           | color of the photons, I wonder how similar this is to
           | camelions and actopi and cuttle fish...
        
             | HarHarVeryFunny wrote:
             | What's most interesting to me isn't just how this plant
             | does it, but how such a capability evolved ...
             | 
             | There has to have been other more primitive sensing/copying
             | capabilities that evolved into this, but what are/were
             | they, and what were the evolutionary benefits of those more
             | primitive capabilities ?
             | 
             | Are there any other plants still existing today that react
             | to light in more than just a present/not-present or
             | directional way ?
        
               | samstave wrote:
               | I have thought and talked about this so often with
               | people. It is fascinating, such as the plants that mimic
               | the shape of the birds that feed upon them - such as
               | hummingbirds.
               | 
               | But the genes are not spread leaf-to-leaf - its through
               | the roots, and may even be communicated through mycelium
               | highways in the root systems where the same mycelium org
               | is symbiotic with both plants.
               | 
               | I wonder what cultivating and formative impact "farming
               | mycelium" might cause in an eco system.
               | 
               | I wonder if the plant in question is able to decode genes
               | that it receives via the Hyphae fungal network, and knows
               | the genes for _" aesthetic and functional appearance for
               | survival here, based on the active genes from the
               | successful plants around me>"_ and can just sideload
               | those genes into its makeup.
               | 
               | Does its offspring, or even clones - removed from that
               | mimic'd env also continue to look like the mimic'd plant?
               | 
               | --
               | 
               | I am pretty sure that the transfer of genes is from the
               | horizontal gene transfer among "neighboring leaves" as it
               | were - but perhaps among the root, as shared by mycelium
               | (fungi) rhizal -- mycorrhizal root structure which shares
               | nutrients between a plant and the fungus.
               | 
               | > _mycorrhizae, known as root fungi, form symbiotic
               | associations with plant roots. In these associations, the
               | fungi are actually integrated into the physical structure
               | of the root. The fungi colonize the living root tissue
               | during active plant growth. Through mycorrhization, the
               | plant obtains phosphate and other minerals, such as zinc
               | and copper, from the soil. The fungus obtains nutrients,
               | such as sugars, from the plant root. Mycorrhizae help
               | increase the surface area of the plant root system
               | because hyphae, which are narrow, can spread beyond the
               | nutrient depletion zone._ <-- may lead to why the plants
               | may seem far enough apart.
               | 
               | https://i.imgur.com/s3Xv6TF.png
               | 
               | https://i.imgur.com/PVn6grj.png
               | 
               | https://apsjournals.apsnet.org/doi/pdf/10.1094/MPMI-05-14
               | -01...
               | 
               | https://bio.libretexts.org/Bookshelves/Introductory_and_G
               | ene...
               | 
               | --
               | 
               | Here is another video on these plants though:
               | 
               | https://www.youtube.com/watch?v=5UWWsYw9fH0
               | 
               | (this guy is talking about bacteria, but its fungi)
        
               | HarHarVeryFunny wrote:
               | If a single plant can simultaneously mimic two different
               | type of neighboring plant, then it seems unlikely that
               | any potential horizontal gene transfer is happening
               | through the roots, where it might be expected to affect
               | the entire plant, not just one localized side.
        
         | throwawaaarrgh wrote:
         | Wouldn't need to be visual. Just requires some way of
         | determining the placement of specific compounds. In most
         | organisms this is by generic coding; your cells don't see a
         | human standing next to your embryo and then decide to make you
         | look like a human. If the plants aren't connected it could be
         | inferred through other means, electromagnetic, auditory, maybe
         | some plant cells move through the air.
        
           | p1mrx wrote:
           | Those non-visual mechanisms do not apply to a plastic plant.
        
       | NotGMan wrote:
       | The book "The Secret Life of Plants" is a very interesting ones
       | and sheds interesting light on many unknown features of plants.
       | 
       | But it requires an open mind, not something most acedemics are
       | known for :)
        
         | block_dagger wrote:
         | Skepticism is a valuable pillar of science. Extraordinary
         | claims, etc.
        
       | jemmyw wrote:
       | I looked at the photos on the article and I'm hard pressed to
       | call that mimicry. The leaves of the vine could just have a
       | ragged and smooth configuration.
       | 
       | The Wikipedia entry for this plant just seems to repeat the same
       | source as the article. So colour me skeptical for the moment.
        
       ___________________________________________________________________
       (page generated 2023-12-24 23:00 UTC)