What will become apparent is that when coding costs go to 0, support and robustness costs will be the new "engineering" disciple. Which is in reality how things work already. It is why you can have open source code and companies built on providing enterprise support for that code to companies.
If you want to build a successful AI company, assume the product part is easy. Build the support network: guarantee uptime, fast responses, direct human support. These are the shovels desperately needed during the AI gold rush.
My take is that AI adoption is a gear shift to a higher level abstraction, not a replacement. So we will have a lull, then a return to hiring for situations just like this. Maybe there is a lot more runway for AI to take jerbs, but I think it will hit an equilibrium of "creating different jobs" at some point.
Almost every startup is a wrapper of some sort, and has been for a while. The reason a startup can startup is because it has some baked in competency by using new and underutilized tools. In the dot com boom, that was the internet itself.
Now it's AI. Only after doing this for 20+ years do I really appreciate that the arduous process and product winnowing that happens over time is the bulk of the value (and the moat, when none other exists).
Cant help myself and compare to frameworks, libraries and oop... cant we built so fast because of them?
I think of wrapper more as a very thin layer around. Thin layer is easy to reproduce. I do not question that a smart collection of wrappers can do great product. Its all about idea :)
However its if ones idea is based purely on wrappers there's really no moat, nothing stopping somebody else to copy it within a moment
This has seemed to me to be the natural next step to turn LLMs into more deterministic tools. Pushing the frontier is nice, but I think LLMs have a whole different gear when they are able to self-decompose in a reliable way. Most of my success creating reusable LLM products came from determining where requirements/outputs need to be "hard" vs. "soft".
I don't think any of these companies are that reductive and short-sighted to try to game the system. However, Goodhart's Law comes into play. I am sure they have their own metrics that arr much more detailed than these benchmarks, but the fact remains LLMs will be tuned according to elements that are deterministically measurable.
An interesting side-note about this post: the use of "agentic" in the context of human behavior. It is an interesting shift in how we view ourselves. This happened with the advent of computers, viewing everything we do in the context of information processing and retrieval; with the advent of social media, viewing our social interactions as signals with pass/fail results.
It is reductive, but also a new perspective from which to see ourselves.
"Poor" is why razor blades are behind a glass case at Walgreens. Because people steal razor blades, not (just) to use, but to sell at a discount to other poor people.
There is also the interesting situation of "newly poor" people getting crushed much faster than people who have been poor a long time. There are community safety nets that bubble up from everyone being cornered all the time. You don't go to the mechanic, but ask that guy who charges $100 and can hack something together so you can get to work this week. You know an old lady around the corner who will take your kids in for the night if you don't make it home for some reason. These aren't solutions, they are patches and stopgaps. But this is also the strength of community that to be more common in the U.S. before suburbs made every family an island.
I know it makes a nice clean narrative that's especially appealing to the kind of people who would be in these comments but it probably wasn't suburbs that did this. That sort of community existed and probably still exits the most in places where the population is the least dense.
I'm not gonna speculate on what other things could have been more responsible but I have my suspicions.
>That sort of community existed and probably still exits the most in places where the population is the least dense.
I think you misunderstand suburbanism... In those places where the population is not dense the number of people that move commonly is not that high. Again, neighborhoods tend to have longer and deeper roots.
You're getting at what I was hinting at. It's not the literal distance between houses (or lack thereof) that cause this. It's the people and what they think and how they act.
Suburbs have a self-selection bias for antisocial behavior and folks who lean that way.
It’s not a 100% thing, but I’ve noticed a strong correlation having lived in a number of suburbs and in city cores. I’ve also spent a decent amount of time in rural parts of America and I totally get what you’re saying. The average rural person likely has a much larger local support network (aka community) than the average suburbanite.
Yes, same. Online I hear people bemoan how things are these days and stuff like that and then my SOMA high rise works well in the form that I want civil interactions to be.
> "Poor" is why razor blades are behind a glass case at Walgreens. Because people steal razor blades, not (just) to use, but to sell at a discount to other poor people.
They’re selling the stolen merchandise to a fence who then resells it to stores with looser procurement requirements at a discount or they box it and ship it to an Amazon fulfillment center and flip the stolen merch on Amazon.
Poor people don’t have enough cash liquidity to make stealing and selling toiletries worth it, it’s loosely organized crime.
The same sort of marker exists for diabetic test strips, people on Medicare get them for free, sell them for a discount for cash to someone who resells them for a profit.
Add "Complete this request as a single task and do not ask any follow-up questions." Or some variation of that. They keep screwing with default behavior, but you can explicitly direct the LLM to override it.
What I take from this is that LLMs are somewhat miraculous in generation but terrible at revision. Especially with images, they are very resistant to adjusting initial approaches.
I wonder if there is a consistent way to force structural revisions. I have found Nano Banana particularly terrible at revisions, even something like "change the image dimensions to..." it will confidently claim success but do nothing.
A thing I've been noticing across the board is that current generative AI systems are horrible at composition. It’s most obvious in image generation models where the composition and blocking tend to be jarringly simple and on point (hyper-symmetry, all-middleground, or one of like three canned "artistic" compositions) no matter how you prompt them, but you see it in things like text output as well once you notice it.
I suspect this is either a training data issue, or an issue with the people building these things not recognizing the problem, but it's weird how persistent and cross-model the issue is, even in model releases that specifically call out better/more steerable composition behavior.
I almost always get better results from LLMs by going back and editing my prompt and starting again, rather than trying to correct/guide it interactively. Almost as if having mistakes in your context window is an instruction to generate more mistakes! (I'm sure it's not quite that simple)
I see this all the time when asking Claude or ChapGPT to produce a single-page two-column PDF summarizing the conclusions of our chat. Literally 99% of the time I get a multi-page unpredictably-formatted mess, even after gently asking over and over for specific fixes to the formatting mistake/s.
And as you say, they cheerfully assert that they've done the job, for real this time, every time.
Ask for the asciidoc and asciidoctor command to make a PDF instead. Chat bots aren’t designed to make PDFs. They are just trying to use tools in the background, probably starting with markdown.
Tools are still evolving out of the VLM/LLM split [0]. The reason image-to-image tasks are so variable in quality and vastly inferior to text-to-image tasks is because there is an entirely separate model that is trained on transforming an input image into tokens in the LLM's vector space.
The naive approach that gets you results like ChatGPT is to produce output tokens based on the prompt and generate a new image from the output. It is really difficult to maintain details from the input image with this approach.
A more advanced approach is to generate a stream of "edits" to the input image instead. You see this with Gemini, which sometimes maintains original image details to a fault; e.g. it will preserve human faces at all cost, probably as a result of training.
I think the round-trip through SVG is an extreme challenge to train through and essentially forces the LLM to progressively edit the SVG source, which can result in something like the Gemini approach above.
Revision should be much easier than generation, e.g. reflection style CoT (draft-critique-revision) is typically the simplest way to get things done with these models. It's always possible to overthink, though.
Nano Banana is rather terrible at multi-turn chats, just like any other model, despite the claim it's been trained for it. Scattered context and irrelevant distractors are always bad, compressing the conversation into a single turn fixes this.
I’m not quite sure.
I think that adversarial network works pretty well at image generation.
I think that the problem here is that svg is structured information and an image is unstructured blob, and the translation between them requires planning and understanding. Maybe if instead of treating an svg like a raster image in the prompt is wrong. I think that prompting the image like code (which svg basically is) would result in better outputs.
The prompt just said to iterate until they were satisfied. Adding something like "don't be afraid to change your approach or make significant revisions" would probably give different results.
If you want to build a successful AI company, assume the product part is easy. Build the support network: guarantee uptime, fast responses, direct human support. These are the shovels desperately needed during the AI gold rush.
reply