Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Apps SDK

developers.openai.com

Post reply on HN
Print view View original

321–330 of 397 posts

Re: Apps SDK

#321

Post by ares623 »

selvan wrote: post #308
Illniyar wrote: post #181

I can't understand the documentation. How are the interactive elements embedded in the chat? Are they just iFrames? The docs mention returning resources, and the example is returning a rust file as a resource, which is nonsensical. This seems similar to MCP UI in result but it's not clear how it works internally.

An MCP server exposes tools that a model can call during a conversation and returns results according to the tool contracts. Those results can include extra metadata—such as inline HTML—that the Apps SDK uses to render rich UI components (widgets) alongside assistant messages. More: https://github.com/openai/openai-apps-sdk-examples?tab=readm...

Imagine rendering content from an app with user submitted data.

Re: Apps SDK

#322

Post by gwd »

esafak wrote: post #55
rushingcreek wrote: post #6

I think this is very interesting, but it is reminiscent of what we built with Phind 2 where the answer could include dynamic, pre-built widgets. The problem with this approach is precisely that these apps/widgets have hard-coded input and output schema. They can work quite well when the user asks something within the widget's capabilities, but the brittleness of this approach starts showing quickly in real-world use.…

The problem is not the limitations of the capabilities per se but their discoverability ( https://en.wikipedia.org/wiki/Discoverability ). The user doesn't know what the capabilities are, as they are added and -- infuriatingly -- removed. Google Assistant is a perfect example of this. Conservational user interfaces are opaque; they lack affordances. https://en.wikipedia.org/wiki/Affordance

Voice interfaces actually remind me a lot of command-line interfaces: If you know the a working "rune" on the tip of your tongue (e.g., "Set a timer for 10 mintues", "Play ") it's great. But as you say, it's not always that easy to figure out new "runes". LLMs should be somewhat better for that, though.

Re: Apps SDK

#324

Post by falcor84 »

totallymike wrote: post #269
sert_121 wrote: post #246

It's interesting to see how chatgpt is becoming more and more of a starting point of the web exploration, at which they're like, why even bother searching at this point, we'll just have default workflows for maps, buy (integration of stripe already marks it), booking airlines etc, which covers so much basic stuff people would do anyways. The biggest bottleneck for this for the past two years imo wasn't the models, bu…

I dismay at the possibility of this happening. What’s the point of an internet at all if one company controls, filters, and governs our entire usage of it? I understand an argument can be made that google is doing similar, but at least you can still search and end up on an actual site, rather than just play telephone via chatgpt. This concept is horrifying for so many reasons.

> What’s the point of an internet at all

Going back to first principles, we need to recall that the internet is for the dissemination of cat pictures, and at the end of the day every technical and organizational change must be analyzed through the lens of its impact on the effective throughput of these pictures.

Re: Apps SDK

#326

Post by willtemperley »

darkamaul wrote: post #304
ukFxqnLa2sBSBf6 wrote: post #277

Earlier quoted context omitted.

I’m not sure how many people there are like me outside of this website but there’s not a single bone in my body that wants to use AI for these things. Buying plane tickets for example. It’s not even that I don’t trust the AI or that I’m afraid it might make a mistake. I just inherently want to feel like I’m in control of these processes. It’s the same reason I’m more afraid of flying than driving despite flying being…

I don’t see why you wouldn’t book a flight using an AI assistant. No one’s saying it should do it completely unsupervised (maybe that’ll come much later), but having something that can research the best routes based on my criteria and show me several options — with a single click to purchase the one I find most convenient — is something I’d love. It could even work against the dynamic pricing algorithms airlines use…

There's probably little danger to the savvy user who understands how manipulative technology like this can be.

The problems come when vulnerable users are targeted using dark patterns. How AI dark patterns will evolve is very uncertain [1] however I suspect they will be extremely subtle and very effective.

What's the worst that can happen if someone vulnerable is persuaded to buy a flight by an AI. I don't know, maybe depression and bad credit after the chatbot's promises weren't met. If they're persuaded to buy a weapon, that's a different matter.

At least current advertising is somewhat public, although that's increasingly less true as ads get more targeted.

This is new territory where ads will be so extremely private it will be only known by the user (maybe they won't even notice) and someone reading the subpoenaed chat logs after a user does something terrible. Those chat logs will likely be inconclusive anyway.

[1] https://venturebeat.com/ai/darkness-rising-the-hidden-danger...

Re: Apps SDK

#327

Post by zengineer »

darkamaul wrote: post #304
ukFxqnLa2sBSBf6 wrote: post #277

Earlier quoted context omitted.

I’m not sure how many people there are like me outside of this website but there’s not a single bone in my body that wants to use AI for these things. Buying plane tickets for example. It’s not even that I don’t trust the AI or that I’m afraid it might make a mistake. I just inherently want to feel like I’m in control of these processes. It’s the same reason I’m more afraid of flying than driving despite flying being…

I don’t see why you wouldn’t book a flight using an AI assistant. No one’s saying it should do it completely unsupervised (maybe that’ll come much later), but having something that can research the best routes based on my criteria and show me several options — with a single click to purchase the one I find most convenient — is something I’d love. It could even work against the dynamic pricing algorithms airlines use…

I would try using AI to book flights - then double check if I can't get a better offer. Do this a couple of times and when I see AI is as good or even better at getting me flights, then sure, why not use it.

Re: Apps SDK

#328

Post by quitit »

OccamsMirror wrote: post #293
tokioyoyo wrote: post #282

Earlier quoted context omitted.

I have very normie and not-so-normie friends that ask ChatGPT almost anything. My parents consistently make use of it, and they're almost 70, and not that tech-literate. There was a fun release from Anthropic about the type of queries that they're receiving, and code-gen is minority. I think we're, once again, not the average user.

I wonder how many of those “average users” will actually happily pay what the true cost is though? Are they really getting perceivable value for it or is it just more convenient than present day Google.

When you say "true cost" I've interpreted that to mean the down-sides of using an AI ChatBot as a primary information source, so my reply below is in context of that interpretation:

There is a sizeable chunk of people who (perhaps foolishly) trust ChatGPT despite knowing it can produce errors. They use it because it does the "research" for them, and does so quickly. This presents a type of tech-agility that they themselves do not possess. So on balance they may be more tech-empowered by using a flawed AI ChatBot than they are by manually reading news, websites and blogs.

There is also an issue of trust. A novice reading the top 5 search results has no real idea if the information being presented is biased, error free, or even factual. Google's work to blend paid and organic placement also presents the flaw of dollars over quality. ChatGPT on the other hand presents a known level of trust to them.

A similar scenario plays out with the way novices are more trusting of apps that appear on a curated store rather than seeking out software via web searches.

I think that users on HN take for granted that they have outsized experience and skill in developing trust in the tech landscape, and have a mental list of news, websites and software providers that they deem trustworthy. This can lead to not understanding the motivation for relying on an AI ChatBot, or compartmentalising people who use those services as some kind simpleton.

Re: Apps SDK

#329

Post by whazor »

heavyset_go wrote: post #270
sert_121 wrote: post #246

It's interesting to see how chatgpt is becoming more and more of a starting point of the web exploration, at which they're like, why even bother searching at this point, we'll just have default workflows for maps, buy (integration of stripe already marks it), booking airlines etc, which covers so much basic stuff people would do anyways. The biggest bottleneck for this for the past two years imo wasn't the models, bu…

Just like I won't trust voice assistants to make purchases for me, or make decisions that actually matter, there is no way in hell that I'm letting an LLM be able to charge my credit card, let alone book flights for me lol

As I understand it, ChatGPT loads the app, performs safe actions, but in the end shows you an UI to confirm the purchase.

Re: Apps SDK

#330

Post by alganet »

anal_reactor wrote: post #222
alganet wrote: post #198

Earlier quoted context omitted.

> Of course the problem is that most chatbots aren't smart. But this is a purely technical problem that can be solved within foreseeable future. Ah yes, it's just a small detail. Don't worry about it.

I don't understand how come that a website for tech people turned into a boomerland of people who pride themselves in not using technology. It's like those people who refuse to use computers because they prefer doing everything the old-fashioned way and they insist on the society following them.

Maybe you can have discussions with a chatbot instead. They always agree with you.
Post reply on HN