Live data from Hacker News

AI and the Automation of Work

ben-evans.com

261–268 of 268 posts

Re: AI and the Automation of Work

#261

No one who claims to automate testing ever seems to have a sophisticated understanding of testing. I suspect the urge people have for automating anything comes not from the love and skill they possess for that activity, but rather for the imagined fruits of it. Software testing is easy to fake, which means it’s particularly easy to claim automation accomplished. Therefore, all these people thinking they are going to…

What's the gap between a Co-pilot written test and a human-written test? Are tests written by co-pilot better than no tests? Developers generally hate writing tests, so I suspect more and more unit tests will be generated. I've worked with someone who does this and it seems insane to me, although I never got to see the generated tests they were referring to, so I can't give specifics.

The first fallacy is that you can't "write" a test, because a test is an event, not an artifact. What coders mean when they say they "write a test" is that they have written an output check. You can write a test procedure, but the test procedure is not the test itself. If you sit down and watch a tester in the act of testing (try videoing it and analyzing 2 sec. slices, while labeling everything that is done, as I have) you will see that testing is a very rich process. It has a heuristic, rather than an algorithmic nature. It involves substantial social competence and judgment, which is a tacit skill set. It is not encodeable.

The next problem is that ChatGPT is not thinking critically about risk. It thinks (well, no it doesn't think... let's say it "treats")... it treats testing as a process of demonstration. Demonstration is not testing. Demonstration proves that functionality is possible, but not that it is reliable. Thus, it produces checking code that is shallow.

The next problem is how are you going to explain to ChatGPT what your product does? Do you give it all your source code? Do you give it all your Jira tickets? The demos I have seen are toy examples. Nothing on a realistic scale.

Let's say you feed it a whole spec and source code, somehow. The next problem is omissions. ChatGPT arbitrarily stops producing output. You then have to carefully check its work to see what it has left out.

What you are asking ChatGPT to do, really, is to write code that will maximize the probability of spotting a real bug, while minimizing the probability of a false positive. But it will arbitrarily focus on only those kinds of bugs that it can easily discover. This is the oracle problem.

The next problem is that ChatGPT halucinates or misunderstands, so you have to correct its mistakes. Sometimes going through frustrating iterations of prompts, like a man trying to pull a mule through town.

You also have trouble with test data and data setup. Only in toy examples is this not a significant problem.

ChatGPT produces conventional results, and that does have some value. But it's not enough for professional work.

I just think the people who casually say that ChatGPT is going to easily handle these things are pretty stupid, under the definition of stupidity as "the refusal to think."

Re: AI and the Automation of Work

#262

Earlier quoted context omitted.

Not really afraid of that at all. On the contrary, you cannot just take an AI and make it do stuff. You need experts to parameterize and train it. It is also a bit more complicated than a pid controller where you turn a few knobs. For training you would need domain experts. The same reason we didn't see too widespread automation in industry. You actually need a lot of engineers to automate systems for their specific…

I agree that engineers will likely be fine (if at all decent). But there are many more jobs done on computers that are fairly routine which are the ones that will be threatened. > An AI currently isn't even capable of doing excel business analysis without the data being checked. This check has to be done by an analyst as well. It cannot even do book keeping without errors and the work to check for errors surpasses do…

>I agree that engineers will likely be fine (if at all decent). But there are many more jobs done on computers that are fairly routine which are the ones that will be threatened.

for instance, transcription?

Re: AI and the Automation of Work

#263

I had bought into the GPT-4 hype. I used GPT-4 on ChatGPT Plus for two months, but then I unsubscribed because I wasn't using it as much as I thought I would. It's certainly a very interesting technology. But in it's current form, I just don't see it replacing most jobs. Will it replace jobs if we scale further? I don't know, but my intuition says no. I feel that we will need some other major breakthroughs in AI to g…

The combination of ChatGPT and Phind helped me, a noob with 8 months of coding knowledge, build an app with real paying users and revenue. I would have never accomplished this on my own, and now I have a working prototype to find co-founders and possible investors.

may i ask how you got paying users?

Re: AI and the Automation of Work

#264

Earlier quoted context omitted.

Would you mind sharing the app link? Curious to see what you made with the help of AI.

Unfortunately that would mean doxxing myself. But its largely a somewhat complex CRUD app with a user dashboard. The stac is nextjs, tailwind, and nodejs, express on the backend.

Did you use templates for the dashboard or just GPT, any complex prompting resources?

Re: AI and the Automation of Work

#265

Earlier quoted context omitted.

Yea, I don't think this will happen over the longterm. I do see worker dislocation in certain sectors though, can't see how LLM's won't substantially shake-up office work especially as Microsoft is already busy building them into the Office Suite.

>I do see worker dislocation in certain sectors… Which was my point. Knowledge workers (and the types of people who gravitate to knowledge work careers) will be dislocated and what work remains may not be as rewarding both financially and personally to these kind of people. People of course adapt, but would that person sitting on the autism spectrum who can thrive and make exceptional income as a SWE be able to adapt…

I would heavily recommend becoming an SWE if they have an inclination for it. It won't be deprecated by AI any time soon and there is still so much work...

AI will become a tool instead of it replacing SWEs. Chances are that software engineer even become more essential through that if AI is integrated in every system.

Re: AI and the Automation of Work

#266

Earlier quoted context omitted.

Unfortunately that would mean doxxing myself. But its largely a somewhat complex CRUD app with a user dashboard. The stac is nextjs, tailwind, and nodejs, express on the backend.

Did you use templates for the dashboard or just GPT, any complex prompting resources?

I built most off my UI off of FlowBite (which is a tailwind UI library). Did not need to use complex prompting resources, though I would sometimes switch between Phind and chatGPT-4 to get a better answer.

I think it works because its not a very complex app and since it's gated with a sufficiently high entry, I don't expect a ton of traffic, hence scaling is not an issue. I also chose the easiest options to deploy (Vercel for the frontned, Heroku for the backend), which might end up costing me in the long run - but I figure I'll get that fixed later

Re: AI and the Automation of Work

#267

Earlier quoted context omitted.

>I do see worker dislocation in certain sectors… Which was my point. Knowledge workers (and the types of people who gravitate to knowledge work careers) will be dislocated and what work remains may not be as rewarding both financially and personally to these kind of people. People of course adapt, but would that person sitting on the autism spectrum who can thrive and make exceptional income as a SWE be able to adapt…

I would heavily recommend becoming an SWE if they have an inclination for it. It won't be deprecated by AI any time soon and there is still so much work... AI will become a tool instead of it replacing SWEs. Chances are that software engineer even become more essential through that if AI is integrated in every system.

I disagree, I think SWE will be one of the first careers displaced by AI. There is a whole “jobber” class of lower quality SWEs nowadays who are bright enough to do the job, but just barely. The extra pair of hands we needed to help digitize the world.

Once the bulk of the coding can be offset by AI, you probably just need elite SWEs as the check to the system. The lower quality SWEs that fill that extra pair of hands role that we need now will go away because the code they produce will be of inferior quality to the code that AI will produce. They will be an anti-productive variable to the equation and will be gradually lost to attrition.

Re: AI and the Automation of Work

#268

I am pretty concerned about AGI. I love computer science and programming. I would love to get into research after I finish my Masters in a couple of years. I think it's fair to say that _some day_ we will achieve AGI. It might not be in 10 years, or 1000, but eventually we'll get there. When we do have AGI, what would that mean for someone like me? Would there be any reason to have programmers (or similar) at all? Wo…

The thing that makes me more optimistic about this specific question is chess. Chess has been, from a human perspective, fully solved by AI. But it has never been more popular. Why? I find it both hard to explain and kind of intuitive. Even though we can't beat computers anymore, we can still learn from them, but remain human, and so human play remains its own unique thing, but infused with insights from the super hu…

Love this take! Each contributing and sharing our own style in combination with others.
Post reply on HN