I replaced all our blog thumbnails using DALL·E 2
deephaven.io
deephaven.io
Not, yet. While it's cheap relative to stock images, it's time consuming to generate exactly what you want. Prices for stock images will collapse for the common quick to use images but the price for the specialized high end images will hold their value or even increase in value. Those historical and such images will continue to be valuable.
It will be interesting to see if a specialized job will rise where people will get paid to generate just the right image. It might be called "A.I. image artist " This individual will generate an image with an A.I. but use graphic tools to finalize it for use.
You'll find these specialized positions working with AI as well soon.
We can deride it from not being "traditional" engineering, but engineering as defined as "doing the best you can with tools you have", certainly allows for a prompt engineer to be defined..
This is not a definition of engineering anyone sane holds. It's so broad it can be used to define a floor cleaner as an engineer (or, if you want higher tech involved, a floor cleaner with a roomba).
This is not logic, this is travesty.
The goal of OpenAI isn't to build a whole new industry of AI Artists, it's to make the AWS of creativity, which means it has to be so simple that you don't even need anyone who can write a good sentence. Just has to get good results from whatever they type.
I don't think so. People just need the result, so the AI will simply become a tool of the trade and you won't have any more AI image engineers than you have dedicated Photoshop artists right now.
Manipulating an AI prompt to get what you want is also a specialized skill, but may require an order of magnitude or two less training, or obviate the need for the job entirely.
An art director for a campaign wants a set of images created, and a set of stock photos used. They may have a junior person on their team create those images, each of which could take hours to create, or they can produce those images with an AI tool, which might take minutes. Or the director may simply use the tool themselves for a few minutes and then hand it off to someone else to clean up "in post".
And probably still, downstream designers are going to be showing how they can convert DALL·E 2 imagery to polished finals. Especially after reviewing the blog post, it's really clear that if you want things to come together well for a refined corporate environment you'll need someone doing that. "I love the whale imagery, but I don't like the DALL·E 2 look, what can we do about the whale teeth" or whatever will definitely be a thing.
One thing I have wondered: what will these generators do to the corporate art market -- art bought in bulk for a hospital or office space. Will interior design specialists pick up prompt generation as an added skill?
So, counter-intuitively it may strengthen Stock Image Sites value
[1] https://twitter.com/kevin2kelly/status/1551964984325812224
The same prompt with DALL-E gives me proper silhouettes with no watermark.
[1] https://github.com/CompVis/latent-diffusion [2] https://imgur.com/a/8tOI9QU
The minute Dall-E becomes a threat, these sites will enforce Dall-E to retrain from public / creative commons images. But Stock Image sites will and be one-up on Dall-E
It's almost you don't how to play chess or business strategy.
Be sure to pass this along to Microsoft's GitHub Copilot.
So even if you are right, and courts rule that training is not fair use, it seems likely that all the big tech companies would lobby Congress to restore the prior state of affairs, as not having an ML industry is somewhat of a national security issue if the rest of the world is going full steam ahead on making Skynet.
I swear, world politics and economics isn’t much different from that, intellectually….
Since there is no legal precedent and the law itself isn't clear about this use case, it's basically a huge gamble on a legal gray area at this point. For VCs the risk doesn't really matter as ML startups only need to exist long enough to provide an exit with high ROI and for enterprise companies it doesn't matter as ML products are just one of many ventures for them.
It's worth noting that unlike Germany, where book and newspaper publishers have won rather unusual copyright claims against companies like Google, in the US the big publishing industries to worry about are movies and music, and most ML projects right now seem to focus on generating images or text rather than music or video. If "AI generated music" caught on like DALL-E 2 did, I think we'd see a lot more contention over how copyright law applies to ML training data.
So we could actually follow copyright laws and still have an ML industry. But I’m not sure big tech wants to ask the public for consent. They would rather have free reign to do whatever they want.
By the way I don’t like copyright law or the concept of IP. But I find it a little annoying that I’m supposed to respect IP law and ML stuff can just ignore it. Also if big tech was forced to encourage people to share stuff with an open license, this would be a huge net good for society! Instead nothing changes but big tech gets to take advantage of peoples copyrighted works and artists can’t do anything to stop it. That kinda sucks.
Today’s ML output throws everything into a mixer and then blanket-calls it ML-generated output, because the original training content has been separated from the social and legal frameworks that govern it.
A NN producing a copyrighted work without respecting the license is not fair use, we know that much.
The funniest ones will be where ML independently reproduces a picture from unrelated materials with humans making a futile effort trying to figure out how it obtained this result.
Apparently using copyrighted training data is okay in EU: Directive (EU) 2019/790 … on copyright and related rights Article 4 [1]
[1] https://eur-lex.europa.eu/legal-content/EN/TXT/HTML/?uri=CEL...
Oh, and if you generate an image with Dall-E and there's a face that is distorted, you can use this tool to restore the facial features. https://arc.tencent.com/en/ai-demos/faceRestoration
Very interesting and promising, thanks! But I always get a 500 error when trying to upload an image... Were you able to make it work past the demo?
What a weird decision to only display prompts in the page title, and nowhere else on the page
This was all true when cars became a thing. What’s the market cap of horse production companies before and after?
My dall-e experience is very limited but looking for the right photo out of many is a very time consuming process, at least at designer level.
It's perfect because:
- The images just need to get across a vibe, they don't need to be perfect
- It's a low-value enough use of images that you'd probably never commission a human artist to do them; instead you'd either use stock photos, or skip having images completely
- The nature of header images for a tech blog tends toward the abstract/surreal, which means it's either hard to find the right stock images, or the ones you do find will be super abstract to the point of being boring
All of these make it a great use of the technology
I can't speak for everyone, but with my own experience of reading (at least partially) a dozen or two technical articles almost every day for many years, pointless media is a hallmark of low quality. these days, I just immediately bail with Ctrl+W as soon as I encounter a twitter-pop-culture meme/gif in the header or anywhere near the top. sure, it does mean I skip the 5 out of 100 that were worth reading, I save a lot of time by skipping the 95 out of 100 that weren't.
“Perfection is achieved, not when there is nothing more to add, but when there is nothing left to take away.” — Antoine de Saint-Exupéry
This comes across as remarkably sad to me. Humor is of questionable value as a baseline?
To each their own, I suppose. I enjoy a technical article with animations that explain things, or a well-placed joke/humorous anecdote/etc. Especially in cases where the humor also supports a point or illustrates something. I tend to remember things better when they have some sort of emotional impact, and out of the emotional impacts possible (sad, angry, etc.) I definitely prefer joy, humor, etc.
All in favour of animations that explain things and definitely enjoy a well-placed joke or anecdote.
tech blogs don't need to follow the same practices as clickbait farms. the audience is very different
the infotainment/clickbait/gossip websites do that because they know their target audience. you don't need to do that on your tech blog
2. It will not affect clickthrough rates of people like you who don't like it
But we are talking here about headers with abstract "art". They are more there for style and "vibe".
Placing a Meme/gif/comic comic above the fold seems to indicate the author was unconfident about the actual content hooking the reader, and they decided to try and hook them with humour or recognisable memes instead. Which is a bad sign in itself.
All that custom abstract art in the header really tells you is the author cares about style/vibe. Which I'd argue is not a bad sign in itself; though perhaps it's a warning sign to quickly check other things, like does the style/vibe match the content you are expecting? Is this just low-effort content to attract newsletter signups?
It's also annoying when the header takes up more than half the screen. Especially more than half of a desktop screen. Phones are somewhat excusable. But I'm not sure there is a correlation between that and bad articles. Caring about style is not the same thing as being good at style.
https://ogp.me/#metadata opengraph requires an image, for example
For example I came up with [0] after writing a draft about an old warrior/mercenary in a fantasy-like setting and then put something into dall-e and built upon that just to get the right "vibe". Or if you're into more "cosmic horror" kind of stuff I generated artworks like these [1] which gave me a lot of inspiration for future short stories I'm planning to draft.
I only spent about $15 so far, and a lot of it was just experimenting with artstyle (mostly to get some interesting discord profile pictures and logos) but I feel like I learned a lot. I can't stress how ridiculously cheap it is for the amount of quality artwork you get out of it.
I dislike super generic stock photo at the beginning of an article. It’s completely pointless, sometimes aesthetically unpleasant, often disconnected with the actual content, and hence a distraction.
If neither you nor the reader cares about the stock photo, why not just forsake the thumbnail or use your website’s logo?
> Blog posts with images get 2.3x more engagement
Perhaps I’m just an outlier; but every time I see a thumbnail that has nothing to do with the title, I scroll past it. It’s a negative for me.
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 could spend all day looking at the output of "impressionist cats" and similar queries.
The images used in the blog linked by OP are okay but stylistically all over the place. OP acknowledges how difficult good prompts are to write. Beyond that, though, you still need to think like an art director and establish a way to set a common style to avoid jarring the readers, and Dalle alone can't do that.
It’s just a matter of time until setting a “consistent art style” becomes a feature of these things.
Similarly, asking the AI to produce multiple views/poses of the same thing will likely become a common feature.
Are training sets prepared with systematic variations in individual axes, as an alternative/addition to tagging each of millions of training images on these axes?
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 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
I would not have any of the ones that I've seen this far on my wall, or as my blog icons.
"Jesus takes a selfie" - https://imgur.com/a/togE2ko
I'm pretty sure this snap was three days after the resurrection.
I suspect that overtime there will be many AIs that will target very specialized functions similar to ARC.
I really, really REALLY don't like this fact and I won't be using or endorsing the technology until it's improved.
I also always hated the "deep dream" pictures of lovecraftian dog horrors
Just like AI generated articles are good enough for 99% of content farms out there.
Until Copilot can make the game you want, you cannot replace developers. And until you think AI is ready to replace artists in general, you won't be able to automate game artists.
That's not to say a game with assets largely drawn by AI, and heavily assisted by Copilot, wouldn't be a cool artistic experiment!
This hasn't put game developers out of a job.
Bad compared with what? They certainly convey a lot more information than a randomly generated gravatar.
The cover image generated for the cosmopolitan cover is stunning at first but after seeing it a few times it begins to feel uncomfortable to look at. The uncanny valley is alive and well in many of these images.
Well they're bad at not looking like AI generated art. It's impressive, but I've yet to come across an example that doesn't look like AI generated art. A few seconds of surface level inspection and you can see the weird AI psychodelic circling effect (no idea what the technical name is - eye-ball-ification?)
To be fair, I think this is because "cyberpunk cityscape" as an artform has become so cliché and generic, it's easy for an AI to copy it!
Here's a couple examples I produced with just a little trial and error. FWIW I have an engineering background and zero design experience.
"Frida Kahlo crossed with Julia Child, 4k realistic, expressive photo, hdr" https://labs.openai.com/s/hvFClrAMCXN6zwqJUJwsmYSB
"John Lennon crossed with Paul McCartney, 4k photograph" https://labs.openai.com/s/lb7qw07tdvRPZ9nmkrCmU0RA
Maybe they're not perfect, but I'm impressed as hell. Exploring what's possible by wording prompts differently feels very much like using a search engine for the first time. Give it a year. This technology is going places.
I thought 2GB ought be enough for everybody.
Maybe part of the reason we are so impressed with those is because they break our perception of reality. It looks like the renaissance statues that are made from marble but looks like cloth.
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 engineers large sums of money to write apps. Even with the retractions we've been seeing in hiring and venture capital, there's just enough easy money still there and the capabilities of code-writing AIs isn't quite there yet.
All we need is a significant market correction and the next generation of AI to wipe out a large swath of tech jobs.
The next step regardless is applying technologies like DALL-E to web design, and for said technology to be widely used, open and affordable. We won't need web designers or even UXD.
Then we won't need as many engineers when AI can solve a lot of common problems in building software. AI can do it better because it won't spend inordinate amounts of time dillydallying over next-gen frameworks, toolchains, and preprocessors. AI won't even have to worry about writing "clean" and maintainable code because those things will no longer matter.
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?
I think that in the very long run programming work will be automated, but by that stage we will either be post-scarcity or reconstituted in computation substrate.
I'm looking for ways to hedge my reliance on my skills.
The 20th time you hear that is when you stop caring.
You still need correct code, and the halting problem says you can't prove whether code does what you want it to. At the end of the day, someone needs to be able to go in and fix shit the AI did wrong, and to do that you need to understand the code the AI wrote.
> You still need correct code, and the halting problem says you can't prove whether code does what you want it to. At the end of the day, someone needs to be able to go in and fix shit the AI did wrong, and to do that you need to understand the code the AI wrote.
This might have been your point, but chances are the "code the AI wrote" will be an unmaintainable mess, so "fixing it" means throwing it away and re-doing it.
There's a reason why the general models aren't being released. The second you look under the hood and start poking the unhappy paths you see that it doesn't understand anything and you're talking to something dumber than a hamster.
I lean towards the latter but with a healthy dose of "it's deeply weird and hard to get anything useful from". But that doesn't make it any less magical.
And no - it's not "intelligent" in any human sense.
But I can't relate to people who pooh-pooh it as if there's nothing exciting happening. Either they are deliberately cultivating a dismissive air, or they are deeply jaded and weary.
EDIT - There's a 3rd option. People are making a rhetorical point because they perceive a need to correct an imbalance in the general mood. This is actually the most likely explanation and is often under-appreciated as motivator in public statements. I've noticed it in myself frequently.
This was true for state of the art in 2010: https://xkcd.com/1425/ today you have a free phone app that does both. Of course it also classifies a spoon as a large breasted robin which is why you need a human in the loop. It's even truer in programming.
Pick any random Jira ticket for a large software project. Could an AI understand and implement that feature into a larger project? Can it correctly deploy it without interruptions to production jobs? Will it correctly implement tests and decent code coverage? If there are regressions will it know how to go in and fix them? If there are bugs reported by users will it be able to investigate and accurately fix them? What about when multiple branches of feature development have to be merged, will it know how to do it correctly? Will it know how to write high performance software or just use shitty random algorithms?
If it can’t do these things AI is basically useless. Because this is basically 90% of software development.
One could think that much of art is just pretty form without sense and that is why DALL-E works.
Most art is not "pretty form without sense". It actually has sense and meaning more often than not, so we can debate what a particular piece "means".
The difference with engineering is that art's meaning is way more subjective, and that if I "miss the point" or simply disagree with the consensus on its meaning, this doesn't make an airplane go down or a nuclear reactor to melt down.
With software it might take days of testing to verify the result, and then repeat that for every iteration. Would be cheaper to build the thing!
Where AI might work is in some restricted subset of software, like a web CRUD app where you say "I want an app that stores a TODO list with dates". With the constraints of it being crud, it just needs to AI the database and arrangement of fields and so on.
The AI is not programming so much as it is choosing which "rails-like scaffolds" to initiate.
The serious engineers are all working on things that go far deeper, and they could never be replaced.
If you want to build a business big enough to be listed on the NASDAQ, you need real developers, and you need to pay them real money.
Most likely, the APP-E or GAME-E, given a prompt generate an application or game, will not generate C++/JavaScript/Swift/Kotlin but directly target the pixel space, running in a 60+ FPS loop a single "function" such as `nextFrame(currentState)`.
It will probably be here in the next few years: write a prompt such as "2D game like Mario but with butterflies" and receive a 2GB blob which opens a window accepting inputs and changing the pixels accordingly. Or, something more serious, a prompt like "invoicing application following the laws of France, Material Design, store the database in AWS using <token>". APP-E or GAME-E doesn't need to totally replace software development, just be good enough to replace in 99% of use cases.
Bugs/tests could probably be solved by some mechanism for injecting localized prompts: given an already generated binary, fine tune it accordingly to a list of other prompts.
As for deployment, it's already pretty much solved with the CI/CD solutions galore all around, not sure why you would need generative statistics for it.
What DALL-E offers is a glimpse of the next 30 years, and probably 99% of the infrastructure required to run it to its full potential is not here yet. Just as in 1992 (3 years after the HTTP proposal, but 2 years before the launch of Netscape) there were only glimpses of what a connected world would look like.
My point was that something like APP-E or GAME-E seems very plausible in the near future and it is more likely to render pixels with the underlying logic encoded in an inscrutable sparse matrix, somewhat the consequence of a beefier DALL-E with regard to the data set, the learning modalities, and the attention span, than to write programs to be compiled/interpreted by any current language stack.
In software, yeah boiler-plate and function-level code-generation... I could also see generating trivial UIs for CRUD apps, or no-code data-pipelines for small businesses... maybe even generating high-level architectures for new services... but we're far off from AI auto-generating code for enterprise applications or foundational services. The differentiation being making changes within an existing complex domain/code-base, in contrast with generating new assets from nothing.
Source: My family owns one of the largest civil engineering firms in my home province.
however, I find that my job (SWE) is about 1% programming and 99% strategizing, designing & communicating.
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.
After experimenting with GitHub Co-Pilot I can see that day being 50% - perhaps even just 25% - as far as it used to feel.
For that scenario to be possible, general AI needs to be developed first.
A huge (and awful) part of software engineering is figuring out what exactly the stakeholders want you to build or fix. Sometimes, they themselves don't even know.
Dealing with ambiguos jira tickets, poorly reported bugs, non-existent requirements, missing or outdated documentation; these are the "common problems" in building software. Current AI technology isn't even close to being able to sort these types of problems today, and it won't be until a monumental breakthrough in the field is achieved.
Generating art is "easy" in the sense that art can't be wrong or right, it just is.
Generating the backend of a streaming platform? I'd like to live long enough to see it.
Yeah, but that part can be learned by anyone without a CS degree.
Perhaps not everything in software can be automated, but I could see a team of 10 programmers be replaced by 1 person (programmer or not) skillful enough to control a bunch of AI software tools.
Ask any creative out there what the hard part of their job is.
I already see a clear path that'd take about 20 years to execute properly. That's assuming low pressure conditions and a very large amount of funding though, both of which aren't typically present in reality.
The result is essentially what GP describes, with a path to AGI in the form of extremely competent tool AI. We're going to hit self-assembling programs before we hit true AGI.
I can't say I'm particularly excited to see such things become reality. Fortunately, humans usually find a way to fuck things up. Our species' collective incompetence is the largest barrier to AGI currently, which may be a blessing in disguise depending on how you look at it.
In a good number of cases it is more difficult to communicate what needs to be built rather than actually building the end product.
The recent work with DALL-E 2 echos a similar problem, coming up with a descriptive prompt can be difficult to do and needs fine tuning to be done. Not unlike trying to communicate with a graphic designer your expected intentions and giving similar works to draw from.
One, coming up with a correct description of a program is what computer programming actually is. Implementation is something we're always looking to do faster, so we can describe more behaviors to the computer.
Two, we're nowhere near the scale of software production which would clear market demand. If everyone who writes code for a living woke up and was ten times as productive, there would be more churn than usual while the talent finds its level, but the result would be ten times as much code per year, not five times and 50% unemployment.
Today I wrote a little bit of code to generate a prefix trie and write it out in a particular format, with some extra attention to getting the formatting 'nice'. This took me about three hours.
It won't be long before something in the nature of Copilot could have gotten this down to, maybe, a half hour for results of the same (minimal, acceptable) quality.
Wonderful! Can't wait, I'll be six times as productive at this kind of task.
This might make it hard, on the margin, for some of the more junior and glue-oriented developers to find work, but I think the main result will be more software gets written, the same way using a structured programming language got people further in a day than writing in assembler did.
However, software development is probably the most thoroughly documented job, the job with the most information online how to do it right, the job with the best available training set. There is a lot of quality code (yes, bad code too), a lot of questions and answers with sample code (stackoverflow...) available. Maybe we've even already written most the software needed in the near future, it's just not available to everyone who needs it (because no one knows all the things out there and also these might be in several pieces in several repos).
Now the one critical thing I think is still needed, based on how actually we create software is an agent that has reasonable memory, that can handle back references (to what it has been told earlier), i.e. one that can handle an iterative conversation instead of a static, one time prompt.
This might be a big leap from where we are now or it may seem like one but AI/ML keeps surprising us for the past decade or so with these leaps. Another thing that may be needed is for it to be able to ask clarification questions (again, as a part of an iterative conversation). I'm not sure about this latter one, but this is definitely how we do the work.
I don't know, after all the predictions about self-driving cars, I'm cautious. Especially considering that back then, it almost seemed obvious that we'd have self-driving cars by now. Cars were certainly capable of driving themselves back in 2016, it just seemed like we needed to iron out a few kinks. How long could that possibly take?
Now, I have no idea when it'll happen.
I'm not necessarily saying that it'll take AI forever to do what humans can do. Rather, I think its very hard to make good predictions with all the hype slightly deceptive marketing.
To most people, the "readiness" of self-driving cars doesn't even come about when they would accomplish parity with human drivers across all situations, because they should be better. And we're not even close to parity with human drivers across a large swath of common situations.
Unless you have links showing otherwise...
Also, if your bar for accidents is "slightly better than a drunk driver and even less accountable", sure, then we can fully deploy that right now. Unfortunately this really isn't sufficient and car companies are fully aware. There's a reason Mercedes made headlines by taking responsibility for the car for ten seconds after disengaging the auto pilot and why they're the only ones who are doing this so far.
We can also talk about AI support service crap with accounts banned/locked without the user having any way to know why or prove he didn't break any terms of the contracts.
An AI artist can get away with a 5% success rate and still be considered viable for replacing humans.
Likewise, and AI programmer can be right only 50% of the time with expert oversight (someone sitting there fixing the code), or 90% with non-expert oversight.
I'm biased from the terrible experience I had trying to get my kids to learn online in the pandemic, but I think schoolteacher might be one of the mass professions that is least susceptible to being AI Engineered away.
Ethicist is probably a safe career path too, but there aren't that many of those. And Politicians will of course prevent robots from taking over their jobs.
At least I can take solace in the fact that for now these things aren’t gonna be taking a bite out of the furry art commissions I like to take, since that’s way too associated with crazy cartoon porn for them to not censor relevant keywords.
“Draw me Elon Musk fellating a Bitcoin in the style of an artist who spent decades refining their craft and producing enough art to train a neural network to make something resembling it, and completely ignore any potential copyright issues because it’s maybe right on the edge of sufficiently transformative, and anyway we have no idea what’s going on in these damn black boxes any more.” That’ll be one compute token, please.
Besides. Even if true, the answer isn't to stop innovating.
In my own life, I've decided to transition away from shared truth towards manifesting the reality that I want to live in. I try to help people now, I meditate a lot about humans becoming aliens, I try to be in the moment whenever possible and be thankful for consciousness. But I no longer put my energy into the ego-based materialism that captured tech. Since wealth inequality can't be stopped, I feel that the only salvation lies in nonattachment.
I also hope this oversaturation doesn’t happen while I’m still alive, or that it happens as quickly as people became obsessed with Mandelbrot sets. You saw them everywhere and then suddenly they weren’t cool any more, and it always feels faintly embarrassing to see one used when you go back to look at something from that era.
I’m fifty years old and I doubt the people making these AI art tools are gonna send a single penny towards supporting and/or retraining the artists they displace, even the ones whose names constantly show up in the lists of things to type into the text box to get a better image.
This insistence bugs the crap out of me, and shows just how arrogant tech-types (and even more so, business-types) can be. The late David Graeber noted this attitude, and while he's talking about investors/entrepreneurs/financiers, it could equally apply to the cocky software engineers I've met over the years:
"It's possible for futurologists to imagine robots relpacing sports editors, sociologists, or real estate agents, for example, yet I have yet to see one suggest that the basic functions that capitalists are supposed to perform, which mainly consist of figuring out the optimal way to invest resources in order to answer current or potential consumer demand, could possibly be performed by a machine. Why not?"
I optimistically believe that art will never be automated away like software engineering has the potential to be. The human creative element is so core to what art is that replacing it with a machine misses the point entirely, even if a machine could fully replicate an aesthetic.
You are not special. I am not special.
Personally my societal concern is yet another industry where we had multiple small jobs will be ruled by a few conglomerates.
Small business are like democracy in a free market. And we keep evolving to it all ending in the hands of a few.
There are already enough offbeat companies dedicated in generating porn, I am sure.
I can‘t wait for a dalle2 level model trained on all of e621, rule34, x-hentai, furaffinity, sofurry, inkbunny and u18chan. It‘s gonna be great.
Here are cups rendered in the style of a famous architect.
https://www.reddit.com/r/dalle2/comments/w7buyx/a_coffee_cup...
Dalle 2's influence will be felt outside the graphic artist realm too.
I mean, I'm guessing these aren't the intended images, since you don't need DALL-E to generate blurry splotches!
I'm not a painter, of course I paint badly. But I find that it is not so difficult to create a picture that expresses my ideas. Mostly I try to use https://excalidraw.com/ for sketching hand-drawn, and find the free stock pictures from https://www.freepik.com/.
I think this is another joy of creation, like programming :-)
Unfortunately the images are cached at a very low resolution
https://deephaven.io/blog/page/2/
etc
I'd prefer smaller thumbnails or icons that give more context to the actual post. This way they could add some benefit, such as helping to visually categorize the content. As of now, they're just a bunch of random illustrations taking up valuable screen real estate.
That being said, thanks for sharing, it's interesting to see an example of someone integrating DALL·E 2 into their workflow.
I see a business opportunity here. Feed text into GPT-3 and have it generate DALL-E prompts to make appropriate images.
Then you have it do the same thing for a children's book.
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.)
Too late. It already started happening a while ago. Tons of blogs with annoying animated gifs and now browsers have the ability to block them.
But it’s also a culture thing…like people trying to convince us that mediocre stuff is good or cool. I mean in this case it’s new because a machine is doing it, but what are OpenAI’s real economic motivations behind all their press releases I wonder…?
Actually, in my experience, and the experience of most of the people in my life, email is a valuable tool that we consistently use, and rarely have any problems with. Spam is a minor inconvenience , at most. (And certainly less of one than the overwhelming number of ads on Facebook.) Phishing is almost non-existent, and very easy to ignore when it pops up a couple of times a year.
In fact, without email, I think the internet would be pretty horrible.
People love to criticize email, but it's actually quite great. It's distributed, low cost, easy to use, and nearly universal. It has mature tools that make the obvious problems (such as spam) not really a problem. And most email senders actually make it easy to unsubscribe. I regularly subscribe/unsubscribe to email lists -- and it's nice having control over what comes into my email inbox.
Email, frankly, is what more of the internet should be like...!
Posts like this are absolutely fascinating!
My prediction is that a lot of blogs will do what you've done too, and then they'll all look/feel the same. New models will come out I guess, but then they'll proliferate too and everything will look the same again. And then maybe to differentiate, those few that value it/can afford it will make the effort to find actually relevant images or commission artwork.
Some less capable algorithms are quite repetitive in such aspect. Perhaps after some time it will be like blog theming where everyone has the same style.
Now that pictures are cheap ven NYT articles are stuffed with photos that don’t add to the article.
Style transfer? It was one of the first CNN applications, and it would be nice to see that included in image generation pipelines.
Think a D&D RPG: I'm writing a bestiary and book of spells, and because I don't care about an exact fit, I can prototype tons of bizarre creatures, wizardly types, weird spells, and colorful scenery. It's pretty cool, and way easier than drawing the prototypes myself.
The one thing these AIs really struggle with is faces. It's hard to get a human without a monstrous face. The best you can hope for is an abstract or blurry face.
And I was thinking exactly this.
This just seems… Not interesting, well done you’ve recreated images that clearly even before AI everyone ignored.
> Use of Images. Subject to your compliance with these terms and our Content Policy, you may use Generations for any legal purpose, including for commercial use. This means you may sell your rights to the Generations you create, incorporate them into works such as books, websites, and presentations, and otherwise commercialize them.
For reference: with cached HTML, my single Node.js process running on Heroku's cheapest tier has weathered the front page multiple times without breaking a sweat
Makes sense for a lot of things, but it comes with downsides, especially for hobbyists! I've found I prefer sticking with a simple server for my website, and OP might find it's easier to do that too
It's just a cache that returns content faster than the original content.
> A content delivery network, or content distribution network, is a geographically distributed network of proxy servers and their data centers. The goal is to provide high availability and performance by distributing the service spatially relative to end users.
Emphasis mine. CDN implies some sort of edge-hosting topology (or at least more-proximal than the main servers).
>CDNs act as trusted overlay networks that offer high-performance delivery of common Web objects, static data, and rich multimedia content by distributing content load among servers that are close to the clients.
and:
>CDNs first emerged in 1998 to address the fact that the Web was not designed to handle large content transmissions over long distances.
[0]: https://ieeexplore.ieee.org/document/1250586
[1]: https://link.springer.com/book/10.1007/978-3-540-77887-5
I think the criteria are that it a) delivers/distributes content b) is a network, implying multiple nodes c) lowers the latency and/or bandwidth cost of data consumption, by d) leveraging geographically distributed redundancy and/or proximity. I think the key feature is geographically distributed redundancy which differentiates it from a regular cache.
OP may or may not feel the same! Just wanted to communicate that a simple server can definitely do the job
1,000,000 uniques over 3 days in ~2011.
I have been trying to re-create that high ever since, lol. Going viral is one hell of a drug.
But as my blog is entirely static (except for the comment threads, hosted on my Discourse forum), I just let CloudFlare serve it. I had to do some tweaks to the configs to say, "No, really, cache everything!" (it doesn't do that by default for a range of very valid reasons, none of which apply to me), but once that change went in, I'll see 98.5% or higher "served out of cache" ratios when I'm seeing a lot of traffic from HN or somewhere.
I'd originally designed it to be hosted out of a Google Cloud bucket with CloudFlare (egress traffic is cheaper that way than out to the internet), but I eventually decided to host on my server, as I could then do Tor and some other stuff more easily. I've got the server anyway...
One of these days, I may play with dropping analytics entirely and just passing requests through to my server, let images remain cached as that's the bulk of my bandwidth. Then I can go even more oldskool and parse my server logs for stats and referrers and such!
Expect to see a bunch of bots. I tried setting up server-side analytics for a WordPress-based website, but I had to get rid of it as the bot traffic made it essentially useless.
If your website is a collection of static files and you're hosting them on S3+CloudFront or something similar (GitHub pages works too), then it'll work without any issues and cost pennies for the whole thing.
Total traffic both times was around 60k over the course of 2-3 days.
https://nginx.org/en/security_advisories.html shows one "medium severity" vulnerability in the last 4 years.
If your goal is to allow people to read the stuff you write, you must be capable of serving your content to >1000 people per second.
Use a good backend (I recommend Servant+Warp or Snap - the only two I've tried that could handle it, out of probably 15 popular options tested) or a good caching reverse proxy (insofar as such a thing can be said to exist) and/or a CDN.