Live data from Hacker News

I replaced all our blog thumbnails using DALL·E 2

deephaven.io

81–90 of 338 posts

Re: I replaced all our blog thumbnails using DALL·E 2

#81
post #9

Am I one of the few people who finds these generated pictures really bad? They often have weird and unsettling details when you look closely. I mean, it’s an incredible achievement in AI that we can generate images at this level, but I don’t want them shown to me on a daily basis while I’m reading blogs.

I tried it a lot and I think it works ok for simple, mundane, trivial prompts, but when you start to ask for sophisticated stuff it gets weird.

Re: I replaced all our blog thumbnails using DALL·E 2

#82
post #9

Am I one of the few people who finds these generated pictures really bad? They often have weird and unsettling details when you look closely. I mean, it’s an incredible achievement in AI that we can generate images at this level, but I don’t want them shown to me on a daily basis while I’m reading blogs.

They are good enough for most people and over time those details will get better until we have no need for illustrators. Already I see website agencies and bloggers using DALL-E. What I do see is that it is easy to pick out DALL-E generated images, in that its too fantastic. Way over the top to a fault.

>Way over the top to a fault.

way over the type as a style

it's like the über-modern modern art. the next level of those goofy over-the-top meme images that make the rounds in socials

while you may not like it, you just know that this will be a thing on how to create AI-like images without AI. I used to refer to that as grade school ;P

Re: I replaced all our blog thumbnails using DALL·E 2

#83

Earlier quoted context omitted.

It's coming for us, too. It won't be long before most software engineer positions are eliminated while some are replaced by software "technicians" with enough expertise to command AI to generate working code. Perhaps the technicians will be tasked with building tests and some automation, but even that stuff can be delegated to AI to an extent. This may seem far off because the present economy is accustomed to paying…

"Physical" engineering fields will probably come first... think AI-generated architecture, with AI-generated structural engineering, plumbing, electrical wiring, etc... with human-guidance of the generative process, and human-review/accountability of final output. Amplification of humans, not obsolescence. In software, yeah boiler-plate and function-level code-generation... I could also see generating trivial UIs for…

Most of the math for structural engineering is already done through software, we just don't call it AI. The difficult part and valuable part of being a good structural engineer is translating requirements and dealing with clients. The actual math and engineering work is often not much more difficult than what's done taught in their undergrad, and much of it is offloaded to designers anyway.

Source: My family owns one of the largest civil engineering firms in my home province.

Re: I replaced all our blog thumbnails using DALL·E 2

#85
post #9

Am I one of the few people who finds these generated pictures really bad? They often have weird and unsettling details when you look closely. I mean, it’s an incredible achievement in AI that we can generate images at this level, but I don’t want them shown to me on a daily basis while I’m reading blogs.

These are good enough for 99% of blogs out there.

Just like AI generated articles are good enough for 99% of content farms out there.

Re: I replaced all our blog thumbnails using DALL·E 2

#86
Ugh, this immediately drives home the realization that representative images are soon going to be devalued and useless, to the point that we'll all be ignoring them soon. Possibly even stripping them with ad blockers or similar tools.

I actually think it's really awesome to be able to do this with a series of blog posts, and even if you look past the stylistic inconsistencies and oddities, this particular usage is good and adds value.

Which is kind of the problem. Relatively low cost, currently high benefit? It's going to be driven into the ground.

We've seen this over and over again. Some reliable form of signal, or of value, becomes inexpensive enough to produce that it gets commoditized, monetized, and weaponized against us all.

Email is a major productivity advance that gives a low-friction way of communicating for mutual gain? Well, now we're drowning in spam and phishing attempts and people won't read random unsolicited messages—if they even make it pass the automated filters. Same for text messages. Bold images and lettering used to be good for highlighting and accentuating important information. Now we don't see them, even if they make it past our ad blockers, because the neural networks living in our skulls know to filter them out as negative-value advertisements.

The same thing will happen here. Nearly all blogs will soon be sprouting cutesy images to go along with the posts. Initially, many of them will be useful and add value, suggesting a metaphor or analogy or simply providing a visual anchor to make the content more memorable.

But they'll quickly become expected and necessary and we'll have the usual race to the bottom. Everyone will have some image because it boosts engagement by 8%... wait now 6%... oops it's too common, we're awash in crappy irrelevant images just added for the boost, which is down to 2%... oh crap, now the absence of an image is a good signal for content quality, we're at -1%!

(If you put work into the prompt and curate carefully, it will still be a net positive to your content. But it won't matter for long in terms of traffic/engagement, because everyone will be mentally ignoring it.)

Re: I replaced all our blog thumbnails using DALL·E 2

#88

When asked about what jobs the robots would come for first, I would have had to say that digital artist was pretty low on my ranking before now.

It's coming for us, too. It won't be long before most software engineer positions are eliminated while some are replaced by software "technicians" with enough expertise to command AI to generate working code. Perhaps the technicians will be tasked with building tests and some automation, but even that stuff can be delegated to AI to an extent. This may seem far off because the present economy is accustomed to paying…

AI is great for recommendation systems and art because they are fuzzy. "Good enough" results are relatively easy to achieve. There is lots of tolerance for errors, because human preferences are flexible and fuzzy themselves.

Engineering is a different ballgame... If anything, all the code monkeys will simply become QA monkeys/test-engineers, because you need to be really sure that your black box algorithm is actually doing what you think it should be doing.

Re: I replaced all our blog thumbnails using DALL·E 2

#89
post #67
post #56

Earlier quoted context omitted.

Would nginx (caching everything) work on a $5/month VPS?

Should work fine. I personally avoided using a reverse proxy like nginx or apache because they tend to have a ton of vulnerabilities (check out the CVE database results for "nginx"), making them a security management headache.

Any serious vulnerability in NGINX will be big news since it is so widespread. CVE database shows some entries by searching for "nginx" but I looked at all 2022 entries and the only ones affecting NGINX itself are in NJX plugin so actually not affecting NGINX core functionality.

https://nginx.org/en/security_advisories.html shows one "medium severity" vulnerability in the last 4 years.

Re: I replaced all our blog thumbnails using DALL·E 2

#90
post #51

Earlier quoted context omitted.

> software "technicians" with enough expertise to command AI to generate working code People keep trying to make simplified programming environments for significantly less-trained people and they keep failing. Is mixing in an AI actually going to make it easier to get a result that has no crippling bugs?

No, but when have crippling bugs ever stopped software businesses from shipping it anyway?

>> People keep trying to make simplified programming environments for significantly less-trained people and they keep failing. Is mixing in an AI actually going to make it easier to get a result that has no crippling bugs?

Yeah. I've even worked in one of those environments for a year (not my choice).

I'm of the opinion those kind of environments won't ever work. They'll either be:

1. Extremely cookie-cutter (e.g. make a clone of our "standard app" with a insignificant little tweaks).

2. Require software engineers to get anything useful out of them, and those engineers will feel like they're working with one hand tied behind their backs (or banging their heads against a wall).

IMHO, one of the main skills of a software engineer is translating user requirements into technical requirements that work and understanding when they work. I don't think skill is automatable without a fairy-tale AGI.

> No, but when have crippling bugs ever stopped software businesses from shipping it anyway?

A lot? Depends on your definition of "crippling." A software engineer will gripe and say, "I don't want to use this;" something that's awkward but the people who use it can still get their work done; or the system literally incapable of performing its function?

Post reply on HN