5 ms·

There is rarely a constructive discussion around the term “AI”. You can’t say anything useful about what it might lead to or how useful it might be, because it

There is rarely a constructive discussion around the term “AI”. You can’t say anything useful about what it might lead to or how useful it might be, because it is purely a marketing term that does not have a specific meaning (neither do both of the words in its abbreviation). Interesting discussions tend to avoid “AI” in favour of specific terms such as “ML”, “LLM”, “GAN”, “stable diffusion”, “chatbot”, “image generation”. These terms refer to specific tech and applications of that tech, and allow to argue about specific consequences for sciences or society (use of ML in biotech vs. proliferation of chatbots). However, certain sub-industries prefer “AI” precisely because it’s so vague, offers seemingly unlimited potential (please give us more investment money/stonks go up), and creates a certain vibe of a conscious being useful when pretending not to be working around IP laws and creating tools based on data obtained without relevant licensing agreements (cf. the countless “if humans have the freedom to read, therefore it’s unfair to restrict the uses of a software tool” fallacies, often perpetuated even by seemingly technically literate people, in pretty much every relevant forum thread).

mnky9800n 1y ago
This article is all about PINNs being overblown. I think it’s a reasonable take. I’ve seen way too many people dump all their eggs in the PINNs basket when there are plenty of options out there. Those options just don’t include a ticket to the hype train.
Closi 1y ago
I think AI is a useful term which usually means a neural network architecture but without specifying the exact architecture. I think Machine Learning doesn't mean this as a word, as it can also refer to linear regression, non-linear optimisation, decision trees, bayesian networks etc. That's not saying that AI isn't abused as a term - but I do think a more general term to describe the latest 5 years advancements in neural networks to solve problems is useful. Particularly as it's not obvious which model architectures would apply to which fields without more work (or even if novel architectures will be required for frontier science applications).
- GrantMoyer 1y ago
  The field of neural network research is known as Deep Learning.
  - danielbln 1y ago
    Eh, not really. All Deep Learning involves neural networks, but not all neural networks are part of deep learning. To be fair, any modern network is also effectively built by deep learning, but your statement as such is inaccurate.
- daveguy 1y ago
  This is incorrect. Machine Learning is a term that refers to numerical as opposed to symbolic AI. ML is a subset of AI as is Symbolic / Logic / Rule based AI (think expert systems). These are all well established terms in the field. Neural Networks include deep learning and LLMs. Most AI has gone the way of ML lately because of the massive numerical processing capabilities available to those techniques. AI is not remotely limited to Neural Networks.
roenxi 1y ago
Also, the strong predictions about AI are using a vague term because the tech often doesn't exist yet. There isn't a chatbot right now that I feel confident can out-perform me at systems design but I'm pretty certain something that can is coming. Odds are also good that in 2-4 years there will be new hotness to replace LLMs that are much more functional (maybe MLLMs, maybe called something else). We can start to predict and respond to their potential even though they don't exist yet; it just takes a little extrapolating. But it doesn't have a name yet. Which is to agree - obviously if people are talking about "AI" they don't want to talk about something that exists right this second. If they did it'd be better to use a precise word.
- Closi 1y ago
  Totally agree. Also the term 'LLM' is more about the mechanics of the thing than what the user gets. LLM is the technology, but some sort of automated artificial intelligence is what people are generally buying. As an example, when people use ChatGPT and get an image back, most don't think "oh, so the LLM called out to a diffusion API?" - they just think "oh chat GPT can give me an image if I give it a prompt". Although again, the term is entirely abused to the extent that washing machines can contain 'AI'. Although just because a term is abused doesn't necessarily mean it's not useful - everything had "Cloud" in it 10 years ago but that term was still useful enough to stick around. Perhaps there is an issue that AI can mean lots of things, but I don't know yet of another term that encapsulates the last 5 years advancements in automated intelligence, and what that technology is likely to be moving forwards, which people will readily recognise. Perhaps we need a new word, but AI has stuck and there isn't a good alternative yet, so is probably here to stay for a bit!
  - giantrobot 1y ago
    > Although again, the term is entirely abused to the extent that washing machines can contain 'AI'. I remember when the exciting term in appliances was "fuzzy logic". As a technology it was just adding some sensors beyond simple timers and thermostats to control things like run time and temperatures of automated washers.
  - strogonoff 1y ago
rickdeckard 1y ago
It's not even that certain sub-industries prefer "AI", it's the umbrella term a company can use in Marketing for virtually any automated process that provides a seemingly subjective result. Case in point: For a decade the implementation of cameras went through development, testing and tuning of Auto Exposure, Auto Focus and Auto White-Balance ("AAA") engines as well as image post-processing. These engines ran on a Image Signal Processor (ISP) or sometimes on the Camera sensor itself, extensive work was done by Engineering Teams on building these models in order to optimize them to run on low-latency on an ISP. Suddenly AI came along and all of these features became "AI features". One company started with "AI assisted Camera" to promote the process everyone was doing all-along. So all had to introduce AI, without any disruptive change in the process.
- Aldipower 1y ago
  I remember somethings similar when the term "cloud" came up. It is still someone else's server or datacenter with tooling.
  - johnisgood 1y ago
    Yeah, can they just stop coining terms to refer to old, pre-existing things? I still hate the term "cloud".
    - idontwantthis 1y ago
      My favorite description is “The cloud is a computer you don’t own in Reston, VA”
- fluidcruft 1y ago
  I agree it's completely meaningless. At this point I think marketing would label a toilet fill valve as "AI".
  - sebastiennight 1y ago
    Well the "smart toilet" is definitely a thing you can buy today: > The integration of Artificial Intelligence (AI) and the Internet of Things (IoT) in bathroom fixtures, particularly toilets, is shaping the future of hygiene, convenience, and sustainability.
jgalt212 1y ago
> There is rarely a constructive discussion around the term “AI”. You hit the nail on the head there. AI, in its broadest terms, exists at the epicenter of hype and emotions.

mnky9800n 1y ago

This article is all about PINNs being overblown. I think it’s a reasonable take. I’ve seen way too many people dump all their eggs in the PINNs basket when there are plenty of options out there. Those options just don’t include a ticket to the hype train.

Closi 1y ago

I think AI is a useful term which usually means a neural network architecture but without specifying the exact architecture. I think Machine Learning doesn't mean this as a word, as it can also refer to linear regression, non-linear optimisation, decision trees, bayesian networks etc. That's not saying that AI isn't abused as a term - but I do think a more general term to describe the latest 5 years advancements in neural networks to solve problems is useful. Particularly as it's not obvious which model architectures would apply to which fields without more work (or even if novel architectures will be required for frontier science applications).

GrantMoyer 1y ago
The field of neural network research is known as Deep Learning.
- danielbln 1y ago
  Eh, not really. All Deep Learning involves neural networks, but not all neural networks are part of deep learning. To be fair, any modern network is also effectively built by deep learning, but your statement as such is inaccurate.
daveguy 1y ago
This is incorrect. Machine Learning is a term that refers to numerical as opposed to symbolic AI. ML is a subset of AI as is Symbolic / Logic / Rule based AI (think expert systems). These are all well established terms in the field. Neural Networks include deep learning and LLMs. Most AI has gone the way of ML lately because of the massive numerical processing capabilities available to those techniques. AI is not remotely limited to Neural Networks.

roenxi 1y ago

Also, the strong predictions about AI are using a vague term because the tech often doesn't exist yet. There isn't a chatbot right now that I feel confident can out-perform me at systems design but I'm pretty certain something that can is coming. Odds are also good that in 2-4 years there will be new hotness to replace LLMs that are much more functional (maybe MLLMs, maybe called something else). We can start to predict and respond to their potential even though they don't exist yet; it just takes a little extrapolating. But it doesn't have a name yet. Which is to agree - obviously if people are talking about "AI" they don't want to talk about something that exists right this second. If they did it'd be better to use a precise word.

Closi 1y ago
Totally agree. Also the term 'LLM' is more about the mechanics of the thing than what the user gets. LLM is the technology, but some sort of automated artificial intelligence is what people are generally buying. As an example, when people use ChatGPT and get an image back, most don't think "oh, so the LLM called out to a diffusion API?" - they just think "oh chat GPT can give me an image if I give it a prompt". Although again, the term is entirely abused to the extent that washing machines can contain 'AI'. Although just because a term is abused doesn't necessarily mean it's not useful - everything had "Cloud" in it 10 years ago but that term was still useful enough to stick around. Perhaps there is an issue that AI can mean lots of things, but I don't know yet of another term that encapsulates the last 5 years advancements in automated intelligence, and what that technology is likely to be moving forwards, which people will readily recognise. Perhaps we need a new word, but AI has stuck and there isn't a good alternative yet, so is probably here to stay for a bit!
- giantrobot 1y ago
  > Although again, the term is entirely abused to the extent that washing machines can contain 'AI'. I remember when the exciting term in appliances was "fuzzy logic". As a technology it was just adding some sensors beyond simple timers and thermostats to control things like run time and temperatures of automated washers.
- strogonoff 1y ago

rickdeckard 1y ago

It's not even that certain sub-industries prefer "AI", it's the umbrella term a company can use in Marketing for virtually any automated process that provides a seemingly subjective result. Case in point: For a decade the implementation of cameras went through development, testing and tuning of Auto Exposure, Auto Focus and Auto White-Balance ("AAA") engines as well as image post-processing. These engines ran on a Image Signal Processor (ISP) or sometimes on the Camera sensor itself, extensive work was done by Engineering Teams on building these models in order to optimize them to run on low-latency on an ISP. Suddenly AI came along and all of these features became "AI features". One company started with "AI assisted Camera" to promote the process everyone was doing all-along. So all had to introduce AI, without any disruptive change in the process.

Aldipower 1y ago
I remember somethings similar when the term "cloud" came up. It is still someone else's server or datacenter with tooling.
- johnisgood 1y ago
  Yeah, can they just stop coining terms to refer to old, pre-existing things? I still hate the term "cloud".
  - idontwantthis 1y ago
    My favorite description is “The cloud is a computer you don’t own in Reston, VA”
fluidcruft 1y ago
I agree it's completely meaningless. At this point I think marketing would label a toilet fill valve as "AI".
- sebastiennight 1y ago
  Well the "smart toilet" is definitely a thing you can buy today: > The integration of Artificial Intelligence (AI) and the Internet of Things (IoT) in bathroom fixtures, particularly toilets, is shaping the future of hygiene, convenience, and sustainability.

jgalt212 1y ago

> There is rarely a constructive discussion around the term “AI”. You hit the nail on the head there. AI, in its broadest terms, exists at the epicenter of hype and emotions.