Lot of astroturfers, no doubt. Its irritating, its like, no, how about they change their behavior and people stop expecting you to treat their computers as computers, which should be personalizable on your side, not just theirs
Yes, this is not labeling what is an apple and what is a pear. "Annotation" does not do this work justice.
For a coding agent, for example, there is *very detailed* analysis of the turns and ranking of different portions of the conversation.
Adherence or deviation from specific rules matters. Writing quality matters. Expertise in the topic under discussion matters. Having intuition for the tone and beat of a good conversation matters.
Scoring a 15-20 turn conversation can easily take two and a half hours.
Clicking submit does not mean the author is done. Many annotations will be turned back to them by a reviewer to touch up in some way.
This work can be far more mentally taxing than programming, is measured much more by completions more of a timed exercise than SWE.
FWIW, Meta employees would probably make great coding agent conversation annotators. But it is absolutely not SWE and they won't enjoy it (for long).
This sub optimal system is what has (at times) led to massive growth of disruptive technologies and platforms.
People got reasonably mad at Apple’s iron fisted control of the App Store and access to its platforms.
It got to the point that there was political will to force Apple to change. (Sort of, in some circumstances)
But during the golden age of the App Store, when Angry Birds came out—-it wasn’t something people were hemming and hawing about the fairness of distribution platforms it was more like: “wow! iPhone! Check out this game! Everyone is playing this thing, Jack White is playing it. You too can have this game for a few bucks, can you believe this?”
It is only after Apple had ridden the horse a good long while and reaped what came to be outsized profits and using the thing to control competition over an extended period that something really was done about it. (Sort of)
There are many other examples of this kind of thing.
Compensated creative expression has somewhat relied on it in the form of intellectual property, which has built in albeit relatively toothless expiration.
But even IP is under greater threat than ever with gen AI. Just look at what fans are doing with the Star Wars franchise on YouTube right now.
Anyhow, I agree with the system is flawed but I mean to point out that deciding a person or company has reaped enough for their thing is something we do on a case by case basis.
It is not easy to say when to say to someone or something has been compensated “enough.”
This is US centric but a $200 Claude code and $100 codex sub is a vast, vast amount of tokens. Enough to pay for itself many times over. It provides exposure to the very edge of harnesses and experience that is being hired for.
Isn’t there an argument this is possibly the best price to available performance for frontier models? Both due to subsidies and the distance between open and accessible alternatives?
I used Kiro in December and I burnt through 200 eur worth of tokens in a weekend. Ultimately it was money well-spent, but, I think that if you want, you can spend as much compute as you have access to. Will it be efficient use of tokens? Probably not.
From all the data, it looks like the 200usd we pay for monthly usage is subsidised… at break-even pricing … well, that 200 is starting to look like a few thousand.
Very neat. Thank you for sharing! I assume this was one shot as well -what sort of prompt did you use?
I’m sure folks would be interested even in a blog post comparing just this process with different Anthropic models if that’s something you do and need a content idea. :)
Can you make a version of this that is more in the style of "the way things work" the cool inventions book from the 90s with cavepeople and wooly mamoths and that illustration asthetic?
Just as impressive was its ability to publish the source and get the version up on my personal site. That was also a one shot but aided by context and skills I have available for these purposes.
I suspect personal privacy and need to run AI workflows to handle the litany of administration tasks of a household will be what result in regular need for local AI.
Apple is already out front with this on a personal, individual level, but they are not obviously headed toward multiuser/family-level ~biz admin with a persistent server running local LLM.
Yes. I got somewhat stuck on the idea of the sets being kept together, partly because I thought it would be good to pass on used ones with the manuals.
But I found that if the builds are out they will be played with and fall apart and eventually become loose legos and that’s all fine and good.
Loose Legos on the floor making random things is fun. But building with sets and instructions is a different skill set and is entertaining in its own right.
The newer Friends series has a short reward video at the end of some builds which sort of puts the cherry on top of the set builds.
It seems like Apple, of all companies has been about doing more with less memory.
That would seem to have helped in the % of BOM getting wrecked by the supply chain and discipline of developers in its ecosystem.
reply