I'm a designer, but I admit it's really hard to predict what products and UX are going to work in the AI age.
Chatbots have been tried for a decade and were a total failure. With AI, it has become the fastest-growing way to interact with technology, ever.
GPT-5 tried to have an auto router and create abstraction for the user; people hated it. Instinct and Muse are doing the same with a different framing; it's going wild.
There are tons of these little examples where a small difference in framing, a small threshold in tech, and product assumptions completely flip on their heads.
It's amazing to follow but requires you to be way more flexible in your assumptions.
I think devil’s in the details on all of your examples, although I very much agree with your overall point - it’s so nuanced in terms of timing, current capabilities, etc.
Re: chatbots were tried but it’s a wrong model for a deterministic system. It’s the best one for a probabilistic one.
Re: chatGPT auto-router - 1) I suspect explicit model selection matters less in agentic contexts (this thing is coordinating a bunch of other agents for me) vs QA context where you want to make sure you’re getting the best answer. 2) I suspect auto-routing matters less now that models are way better
But I have to admit that in the design field we have a strong tendency to over-rely on "proper rules" and to overgeneralize learnings. (My pet peeve is when B2C patterns are transferred as-is in B2B or prosumers)
AI makes some old UX viable but can also unlock tons of value once you break your assumptions and reluctance. An example I see a lot here and on socials is how much of the code developers actually read nowadays. In the case of Instict, it is how many people actually share their credit card with an unknown startup's IA to buy stuff. I just heard on a podcast that 40% of them actually used their AI to buy something online (I didn't verify the %)
I'm the kind of guy who didn't understand why people use phones so much when a desktop is way more comfortable, so I have to keep track of the crazy ones!
The biggest mistake people make in this space is confusing the limitations of a specific implementation of a technology with fundamental flaws with the entire idea of said technology.
Once you recognize and overcome that thinking, and learn to distinguish between those two things, you can see where things are going.
GPT-5's auto routing was bad, you'd get off the cuff bad answers for hard questions and overthinking on the banal. Also, with a chatbot the idea is that it's a back and forth but 5 minutes of thinking kills that, whereas fire and forget is core to the idea of autonomous agents.
Agents are going to replace almost all the UI in computers. For a trivial example, things like a control panel or settings menu will cease to exist, you'll handle coarse things via conversation ("set volume to 20%"), and UIs will be presented only for things where chat is cumbersome ("the password for the printer is !-*Donkey_Loves_Shrek^^@"). The UI that remains will become heavily decoupled, like supercharged versions of the office ribbons, with the intent that the agent will compose menu fragments and capability widgets dynamically so you only ever need to see controls for the settings that are relevant and that you care about.
Sorry for being dismissive. What I mean is, volume is something people generally don’t think about in percentages, but by feel/ear. It’s also adjusted so often and impulsively that doing it via chat would be extremely slow and unergonomic. So yeah, I don’t think that should happen.
Not saying you won't be able to invoke the volume control widget on a screen, but when harnesses are always listening, models are lower latency and you're interacting with a device from the other side of the room in a lot of cases (even PCs, via 72" wall mounted monitors), it'll be a lot more ergonomic.
There was a study that showed users "felt" faster on a keyboard even while completing tasks faster on a mouse. Similarly, with low latency models, "chromium volume 20%, actually, 10%" will be faster than clicking the sound control icon, finding the chromium volume from the list and reducing it, not to mention you don't have to look away from what you're doing. Direct interaction will probably only be faster if you need extremely precise control, for example, you wouldn't do digital painting with an agent, which seems pretty intuitive.
I mean maybe I‘m an outlier but I always use the physical buttons or knobs to control volume on my devices. I think that is by far the fastest, most exact and overall most user friendly and unobtrusive option, even when you have an agent that‘s always listening (which is a future I don‘t want). The point about being across the room is fair.
About that study: I‘m interested to read it, as it seems to contradict established literature on UX (not to mention lived experience) - do you have a link for that?
Tognazzini, Maximizing Human Performance; Omanson et al. 2010, Comparison of Mouse and Keyboard Efficiency; Card, Moran & Newell 1980, Keystroke-Level Model.
I seemed to recall that Tog made that point back then. Note that both Tog‘s text and Keystroke-Level are from the very early days of GUIs, so, grain of salt.
Omanson, OTOH actually contradicts them a bit: expert users are generally faster with well-learned shortcuts – a widely-held view point and the whole reason for shortcuts.
Essentially. I don't know much Linux, but I've had no problems putting together a capable, well-configured, and secure Linux box by using Claude Code as the CLI 99% of the time. Occasionally it gives me a command that needs to be pasted into a separate terminal window, but for the most part it just works on its own (and boy, does it struggle sometimes.)
A key point is that the box might as well be running BeOS or Windows or HP-UX or OS/360 or something. It doesn't matter much anymore, any more than my choice of programming language does. Requires some mental adjustment.
> requires you to be way more flexible in your assumptions.
This is what happens if you try to read the data like gospel without having a mental model to interpret it... in other words, critical thinking.
Chatbots are a failure because chat capabilities have sucked for most of the past decade and chatbots weren't really capable enough to do anything meaningful. Not because chat is fundamentally bad. Similarly with auto-router.
Also, if you want to "cross the chasm" you cannot measure reactions from current AI users (early adopters) and extrapolate to the rest of the populace. Early adopters are paying you for tokens and try to wieid them like a surgical tool; across the chasm the rest really don't want to know anything about "tokens" or models tiers or "thinking level".
Chatbots also had huge drawbacks that are still present with AI:
- Inability to see what outcomes are possible
- Very constrained environment with only lightly formatted text
- Sequential flow
You can see how the modern AI chat interface tries to overcome these issues. But even then, just the power of LLMs made users adopt this approach massively despite these drawbacks
Hindsight is 20/20, maybe you can accurately predict the inflection point of every modality, but the fact that most AI patterns currently used started as a quick experiment shows that most people can't, and probably shouldn't.
Adrig · · focus · HN ↗
Chatbots have been tried for a decade and were a total failure. With AI, it has become the fastest-growing way to interact with technology, ever.
GPT-5 tried to have an auto router and create abstraction for the user; people hated it. Instinct and Muse are doing the same with a different framing; it's going wild.
There are tons of these little examples where a small difference in framing, a small threshold in tech, and product assumptions completely flip on their heads.
It's amazing to follow but requires you to be way more flexible in your assumptions.
young_mete · · focus · HN ↗
Re: chatbots were tried but it’s a wrong model for a deterministic system. It’s the best one for a probabilistic one.
Re: chatGPT auto-router - 1) I suspect explicit model selection matters less in agentic contexts (this thing is coordinating a bunch of other agents for me) vs QA context where you want to make sure you’re getting the best answer. 2) I suspect auto-routing matters less now that models are way better
Adrig · · focus · HN ↗
But I have to admit that in the design field we have a strong tendency to over-rely on "proper rules" and to overgeneralize learnings. (My pet peeve is when B2C patterns are transferred as-is in B2B or prosumers)
AI makes some old UX viable but can also unlock tons of value once you break your assumptions and reluctance. An example I see a lot here and on socials is how much of the code developers actually read nowadays. In the case of Instict, it is how many people actually share their credit card with an unknown startup's IA to buy stuff. I just heard on a podcast that 40% of them actually used their AI to buy something online (I didn't verify the %)
I'm the kind of guy who didn't understand why people use phones so much when a desktop is way more comfortable, so I have to keep track of the crazy ones!
heaney-555 · · focus · HN ↗
Once you recognize and overcome that thinking, and learn to distinguish between those two things, you can see where things are going.
CuriouslyC · · focus · HN ↗
Agents are going to replace almost all the UI in computers. For a trivial example, things like a control panel or settings menu will cease to exist, you'll handle coarse things via conversation ("set volume to 20%"), and UIs will be presented only for things where chat is cumbersome ("the password for the printer is !-*Donkey_Loves_Shrek^^@"). The UI that remains will become heavily decoupled, like supercharged versions of the office ribbons, with the intent that the agent will compose menu fragments and capability widgets dynamically so you only ever need to see controls for the settings that are relevant and that you care about.
ch_sm · · focus · HN ↗
haha oh god no please don’t.
EDIT:
Sorry for being dismissive. What I mean is, volume is something people generally don’t think about in percentages, but by feel/ear. It’s also adjusted so often and impulsively that doing it via chat would be extremely slow and unergonomic. So yeah, I don’t think that should happen.
CuriouslyC · · focus · HN ↗
There was a study that showed users "felt" faster on a keyboard even while completing tasks faster on a mouse. Similarly, with low latency models, "chromium volume 20%, actually, 10%" will be faster than clicking the sound control icon, finding the chromium volume from the list and reducing it, not to mention you don't have to look away from what you're doing. Direct interaction will probably only be faster if you need extremely precise control, for example, you wouldn't do digital painting with an agent, which seems pretty intuitive.
ch_sm · · focus · HN ↗
About that study: I‘m interested to read it, as it seems to contradict established literature on UX (not to mention lived experience) - do you have a link for that?
tty2 · · focus · HN ↗
[dead]
CuriouslyC · · focus · HN ↗
ch_sm · · focus · HN ↗
I seemed to recall that Tog made that point back then. Note that both Tog‘s text and Keystroke-Level are from the very early days of GUIs, so, grain of salt.
Omanson, OTOH actually contradicts them a bit: expert users are generally faster with well-learned shortcuts – a widely-held view point and the whole reason for shortcuts.
CamperBob2 · · focus · HN ↗
ch_sm · · focus · HN ↗
CamperBob2 · · focus · HN ↗
A key point is that the box might as well be running BeOS or Windows or HP-UX or OS/360 or something. It doesn't matter much anymore, any more than my choice of programming language does. Requires some mental adjustment.
ch_sm · · focus · HN ↗
> don't know much Linux, but I've had no problems putting together a capable, well-configured, and secure Linux box by using Claude Code
How can you be sure of it being well configured and secure? Especially if you don‘t know much about Linux?
CamperBob2 · · focus · HN ↗
In practice, that means I at least stick around and watch when Claude tells me to curl | sudo bash something. :-P
mech422 · · focus · HN ↗
tty2 · · focus · HN ↗
[dead]
ssivark · · focus · HN ↗
This is what happens if you try to read the data like gospel without having a mental model to interpret it... in other words, critical thinking.
Chatbots are a failure because chat capabilities have sucked for most of the past decade and chatbots weren't really capable enough to do anything meaningful. Not because chat is fundamentally bad. Similarly with auto-router.
Also, if you want to "cross the chasm" you cannot measure reactions from current AI users (early adopters) and extrapolate to the rest of the populace. Early adopters are paying you for tokens and try to wieid them like a surgical tool; across the chasm the rest really don't want to know anything about "tokens" or models tiers or "thinking level".
Adrig · · focus · HN ↗
Chatbots also had huge drawbacks that are still present with AI: - Inability to see what outcomes are possible - Very constrained environment with only lightly formatted text - Sequential flow
You can see how the modern AI chat interface tries to overcome these issues. But even then, just the power of LLMs made users adopt this approach massively despite these drawbacks
Hindsight is 20/20, maybe you can accurately predict the inflection point of every modality, but the fact that most AI patterns currently used started as a quick experiment shows that most people can't, and probably shouldn't.