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

What it feels like to work in AI right now

robotic.substack.com

Post reply on HN
Print view View original

571–580 of 768 posts

Re: What it feels like to work in AI right now

#571

Post by alexvoda »

bottlepalm wrote: post #311
riffraff wrote: post #298

Earlier quoted context omitted.

> Again, no bugs, perfect code How do you know? Either you know enough Go to be able to tell, and thus you'd be able to write this yourself, or you don't and thus you can't really judge. I mean, it probably is right, but this "I don't know something, but I trust what the chatbot told me" is what worries me about the rise of the LLMs.

Reading and validating code is way easier than figuring out how to write it in the first place. A lot of programming is the same - functions, conditionals, loops, libraries, etc.. The hard part is mostly semantics across languages. Way easier for GPT to do the first pass, then go through it and read up on the parts you don't understand.

> Reading and validating code is way easier than figuring out how to write it in the first place

Not at all. I would even say it is harder because the code being reviewed is priming you.

All code is a leaky abstraction and if you don't know what to look for you just won't see it.

Re: What it feels like to work in AI right now

#572

Post by golergka »

cgearhart wrote: post #152

From my perspective it’s just _confusing_ to work in AI right now. We have some massive models that are doing some really neat stuff, and apparently hundreds of millions of people are using them—but I keep wondering: to do _what_, exactly? I’m not asking what the models can do, I’m asking what people want the models to do every day, all the time. I’ve been shown some neat pictures people made that they thought were c…

Almost all office work is just reading words in boxes and looking at pictures and then writing words in other boxes.

There's insane potential to automate it. Think b2b, not b2c.

Re: What it feels like to work in AI right now

#573

Post by alexvoda »

riffraff wrote: post #298
jack_pp wrote: post #243

Earlier quoted context omitted.

I thought about learning Go and Fiber to build the backend of a side project of mine and I did but as it goes for any new language / stack I wasn't feeling confident in it. Then ChatGPT came out and I thought what the hell, let's see what all the fuss is about. So I asked it to write me a struct for a table with the "id, name, longitude, latitude, news" columns. That worked well, I was surprised it automatically infe…

> Again, no bugs, perfect code How do you know? Either you know enough Go to be able to tell, and thus you'd be able to write this yourself, or you don't and thus you can't really judge. I mean, it probably is right, but this "I don't know something, but I trust what the chatbot told me" is what worries me about the rise of the LLMs.

> I mean, it probably is right, but this "I don't know something, but I trust what the chatbot told me" is what worries me about the rise of the LLMs.

I share the same worry with regards to the way humans will use AI, along with worries about enabling various antisocial behaviours at scale.

Re: What it feels like to work in AI right now

#574

Post by machiaweliczny »

cgearhart wrote: post #152

From my perspective it’s just _confusing_ to work in AI right now. We have some massive models that are doing some really neat stuff, and apparently hundreds of millions of people are using them—but I keep wondering: to do _what_, exactly? I’m not asking what the models can do, I’m asking what people want the models to do every day, all the time. I’ve been shown some neat pictures people made that they thought were c…

Killer app is that it's glue natural language code. It's not realised yet but it's gonna be BIG. It's simply gonna be copilot for everything.

Re: What it feels like to work in AI right now

#575

Post by shmoogy »

sdenton4 wrote: post #313
uxcolumbo wrote: post #281

Earlier quoted context omitted.

I”m not in ML, so excuse this maybe naive question: > get hundreds of LLMs supervising teams of millions of LLMs What does this mean or what can you do with this setup… do you mean running LLMs in parallel?

Ask the LLM to perform a complex task by splitting it into sub tasks to be performed by other LLM instances, then integrate the results...

Is this something like langchain is working towards?

Re: What it feels like to work in AI right now

#576

Post by alexvoda »

staunton wrote: post #567
alexvoda wrote: post #562

Earlier quoted context omitted.

And what is a business that is fine with occasional (or frequent) bullshit output? Fake-news and spam.

Every business is fine with some frequency of bullshit output at some level. The question is how often exactly it happens and how much harm the bullshit can cause.

My point was that spam is the perfect use case for this tech. Of course there are other possible use cases, but spam and fake news content creation are the perfect fit. AI will enable one to easily clone the writing style of any publication and insert whatever bullshit content and keep up with the publishing cycle with almost zero workforce.

Want a flat-earther version of New York Times (The New York Flat Times)? Done. Want a just slightly insidiously fascist version of NPR? Done. Want a pro-Nato version of RussiaToday (WestRussiaToday)? Done.

And we already know people share stuff without checking for veracity and reliability first.

Re: What it feels like to work in AI right now

#577

Post by numpad0 »

fomine3 wrote: post #264
rvz wrote: post #216

Earlier quoted context omitted.

> Taking into account the downsides it looks like a hype bubble right now to me, and a draw in the long run. There’s just a whole lot of tech people trying to cash in on the hype. Techies will realize that they are just giving ideas to O̶p̶e̶n̶AI.com, Microsoft Word, Google Docs and Notion. It is just the same AI bros re-selling their hallucinating snake oil chatbot that are under a new narrative for AI. There is a r…

English speakers tend to miss a huge use case: Translation.

I do wonder where LLM translator would take us to, considering that Japanese version of Bing Image Creator[0] is still proudly displaying a complete nonsense…

0: https://www.bing.com/create

    作成 *芸術* 開始日
    AI を使用した単語

Re: What it feels like to work in AI right now

#578

Post by acdha »

golergka wrote: post #572
cgearhart wrote: post #152

From my perspective it’s just _confusing_ to work in AI right now. We have some massive models that are doing some really neat stuff, and apparently hundreds of millions of people are using them—but I keep wondering: to do _what_, exactly? I’m not asking what the models can do, I’m asking what people want the models to do every day, all the time. I’ve been shown some neat pictures people made that they thought were c…

Almost all office work is just reading words in boxes and looking at pictures and then writing words in other boxes. There's insane potential to automate it. Think b2b, not b2c.

“Almost all” makes me question how much experience you have doing those jobs. They exist but work which is truly that rote has been getting automated for half a century so what I’ve found is that what you see now tends to deceptively appear simple while hiding judgement, uncertainty, and inference.

That’s especially true for things which involve liability. If Google builds a system to recommend YouTube videos and ads, it’s a win as long their error rate is below a certain level. If it’s your insurance company rejecting claims, however, people can die that way and the lawsuits for breaking contracts or legal standards can far exceed the believed savings.

Re: What it feels like to work in AI right now

#579

Post by ModernMech »

sashank_1509 wrote: post #194

This sub and blog post are all talking about ChatGPT which is no doubt amazing and far ahead of the curve. However I would also point to Metas new vision model: SAM (Segment Anything): it is so far beyond any other vision model, I actually believe vision will be solved in a few years now. People don’t realize that there was an industry of publishing paper with incremental improvements in small datasets in CVPR that h…

> I’ve worked in Tesla Autopilot before and saw it hit a mannequin because we never had mannequin in our dataset before ( we might have had it in the data but it was not a part of our ontology for the network to predict).

Incredible. To be clear, your group and the company you work for saw fit to release your beta-quality autopilot onto public roads, as it was hitting human-looking objects in your test labs. Is that what you're admitting here? Did anyone in your group object to this? Were you personally concerned?

Instead of moving fast and breaking things, what if you had not rushed your autopilot out, but waited on technology to improve to the point that your product wouldn't hit a human-looking object it had never seen before?

Re: What it feels like to work in AI right now

#580

Post by machiaweliczny »

Buttons840 wrote: post #245
woeirua wrote: post #214

I think what’s really depressing here is just how effective scaling seems to be. It just means that any company that’s not willing to pour hundreds of millions of dollars into their AI programs isn’t serious at all and would probably be better off hiring engineers to figure out how to integrate GPTX into their systems than trying to roll their own. I really think we’re going to see a massive collapse of AI/data scien…

My understanding is that transformers are now favored over RNNs because they parallelize better. It's hard to imagine, but I wonder if there's some non-parallelizable machine learning algorithms which might outperform these massive models? It seems improbable, but it's a small hope I've had. The greatest intellects were aware of (ourselves) do not scale very well, and maybe the same will ultimately apply to AI?

I remember seeing some theoretical analysis that compared computing differences between transformers, LSTMs and RNNs and I think that RNNs are theoretically better (can learn more complex functions). Can't find it now.
Post reply on HN