My father still has one of these in orange and white. I remember when I was a little child and he would start it up, I could feel the concussion of the exhaust in my chest.
I've been neck deep in a personal project since January that heavily leverages LLMs for the coding.
Most of my time has been spent fitting abstractions together, trying to find meaningful relationships in a field that is still somewhat ill-defined. I suppose I could have thrown lots of cash at it and had it 'done' in a weekend, but I hate that idea.
As it stands, I know what works and what doesn't (to the degree I can, I'm still learning, and I'll acknowledge I'm not super knowledgeable in most things) but I'm trying to apply what I know to a domain I don't readily understand well.
I've found that value is largely derived from polish and vision.
It's easy to prompt some stuff into existence over a weekend. It is hard to polish it, fix bugs, have tidy UX, and so on. There's this meme going around (maybe from that Silicon Valley show?) where the grey-beard says he is valued for his taste and his conviction in that taste. This is -- fortunately or not -- reality.
Vision and taste won't get you the whole way, but they are a huge part of the equation. This is why Apple, for example, was so successful under Jobs: he had vision, and he had good taste.
I agree, and for those who would counter “just use AI to polish”, those who use AI to avoid doing the work of building something are likewise going to avoid doing the work to polish it, if they even possess the taste required to do so.
I've been working on a framework since the end of January or so. I'm on my 7th draft. As I've gone along, each draft gets markedly smaller. The overlaps between what I'm building and openclaw are significant, but I've realized the elements that make up the system are distinct, small, and modular (by design).
There are only a few primitives:
1. session history
1a. context map + rendered context map (think of a drive partitioning scheme, but for context -- you can specify what goes into each block of context and this gets built before being sent out for inference).
2. agent definition / runtime
3. workflow definition / runtime
4. workflow history
5. runtime history (for all the stuff session and workflow history fail to capture because they are at a lower level in the stack)
That's it. Everything else builds on top of these primitives, including
- memory (a new context block that you add to a context map)
- tool usage (which is a set of hooks on inference return and can optionally send the output straight back for inference -- this is a special case inside the inference loop and so just lives there)
- anything to do with agent operating environment (this is an extension of workflows)
- anything to do with governance/provenance/security (this is an extension of either workflows and/or agent operating environment... I haven't nailed this down yet).
I suppose I should say something about how agents and workflows work together. I've broken up 'what to do' and 'how to think' into the two primitives of 'workflow' and 'agent' respectively. An agent's context map will have a section for system prompt and cognitive prompt, and an agent can 'bind' to a workflow. When bound, the agent has an additional field in their context map that spells out the workflow state the agent is in, the available tools, and state exit criteria. Ideally an agent can bind/unbind from a workflow at will, which means long-running workflows are durable beyond just agent activity. There's some nuance here in how session history from a workflow is stored, and I haven't figured that out yet.
Generally, the idea of a workflow allows you to do things like scheduled tasks, user UI, connectors to a variety of comms interfaces, tasks requiring specific outputs, etc. The primitive lays the foundation for a huge chunk of functionality that openclaw and others expose.
It's been fun reasoning through this, and I'll admit that I've had an awful lot of FOMO in the mean time, as I watch so many other harnesses come online. The majority of them look polished, and are well marketed (as far as AI hype marketing goes). But I've managed to stay the course so far.
I hope you find your ideal fit. These tools have the potential to be very powerful if we can manage to build them well enough.
I wonder if stuffing tool call formatting into an engram layer (see Deepseek's engram paper) that could be swapped at runtime would be a useful solution here.
The idea would be to encode tool calling semantics once on a single layer, and inject as-needed. Harness providers could then give users their bespoke tool calling layer that is injected at model load-time.
Dunno, seems like it might work. I think most open source models can have an engram layer injected (some testing would be required to see where the layer best fits).
The engram idea is actually technically clever but imo sees the solution from a bottom-up approach while Louf's real argument is a top-down view. His solution (declarative specs) solves that by centralizing the spec, making it versioned and composable, independent of any actual model.
Engram layers just move the coordination problem earlier and lock it in. Coordination problems between models & providers would still exist, requiring a layer injection in each open source model and another variant produced for each. Users would still need to chose between "Qwen-8b" and "Qwen-8b-engram" x model families and sizes. Is that cleaner?
The issue with a top-level spec, that I can see, is that models fall back to their training when it comes to tools. This is why I recommended the engram approach, because as far as I can tell the problem is a model problem not a systems problem.
Yes, but it is not a production capacity problem. The constraints on food are mostly in the logistics chain, often having to do with corruption or distribution targets (food goes where the money is), or regulation (did you know that cherry growers in the Upper Midwest are required --_by Federal law_-- to destroy unsold crops?).
A huge amount of food goes to waste simply because of regulation or subsidies, at least within the United States.
Tart cherries are supply-controlled because they are processed into other goods, like pie filling, and can be stored for long duration (multiple seasons). The supply-control regulation is designed to prevent a surplus crop from depressing the market to the point where it's no longer viable to grow tart cherries - reducing future supply, ie. the regulation is designed to provide a consistent, stable supply.
Surplus tart cherry crops are rarely destroyed. In the event of a surplus, they are often exported, diverted to secondary markets, donated, or carried-over into next-season's stock.
Yup. The regulations on food in the US is exactly to make sure the shelves stay stocked no matter what. Without such regulations, you'd experience random items being unavailable and price shocks.
One thing people often don't figure or realize is food takes time to grow. It requires long term thinking to make sure supplies are sufficient. Left to their own devices, farmers will often chase after last season's cash crop. That is bad. It's far better for farmers to stick to more predictable growing and for more dedicated incentives to be issued.
Did you intend to be so insulting, condescending, and dismissive? "Left to their own devices, farmers will often chase after last season's cash crop. That is bad. It's far better for farmers to stick to more predictable growing and for more dedicated incentives to be issued."
I grew up on a farm and lived around farmers. This is my lived experience.
I saw first hand farmers tear up a barley fields to plant wheat when the price got high enough.
Farming is a game of speculation. Planting last year's cash crop can be a successful strategy just like buying APPL today will likely yield good returns. Yet, it's a very hard market to predict with a lot of luck involved. Maybe only a few chase the cash crop and you win big. Maybe everyone does and you lose. Maybe there's a natural or political disaster that pumps up your crop.
There was nothing insulting, condensing, or dismissive about my comment. Highly speculative markets, like food, have booms and busts that can swing wildly. That's bad for something like food. The free market does not work with crops.
I'd argue that this should be refined to something like "farmers that speculate heavily struggle in an under-regulated free market".
Financial stability in highly volatile markets depends on appropriate planning, saving, and distribution. I say this from the investment perspective, but I would venture to guess that it also applies to hard goods like food-stuffs.
The nature of farming is speculation. It's inescapable. In a completely free market there's no way to guarantee success. Even with the best planning and saving you can't know what the rest of the market is doing and because of the long tail, you are locked in to harvesting and selling your crop no matter what.
You can speculate and be the farmer that always plants and grows wheat. You'll see booms and busts based on that. You can also switch up what you are growing based on your best guess about demand. Both strategies can be successful.
Funnily, one way to make farming less risky is a futures contract. And, if you know anything about futures commodity trading you know they are some of the most risky forms of trading.
It's true though, these regulations exists because speculation and profit-chasing in agriculture is what lead to the dust bowl and worsened the great depression. We really, really don't want a repeat of that.
The amazing thing about people failing to learn from history is that everybody thinks they're too smart to (a) learn history or (b) follow rules enacted to prevent the disasters of yesteryear.
Learning from history is important but it’s much more important to do so in an inclusive manner. In fact, inclusive language is more important than anything else.
Could you please stop posting unsubstantive comments and flamebait? You've unfortunately been doing it a lot lately. We've already asked you a whole bunch of times not to do this. Eventually we ban accounts that won't stop.
Sure, but I think you should strive to run your community in a way where you’re policing the “I don’t endorse X, but I don’t understand why more people don’t do X” that this comment espouses https://news.ycombinator.com/item?id=47773488
You’re busy policing this while people are out there saying “Destroy their things and firebomb their houses”. So is it just that I made a mistake in my phrasing? Should I just frame the same comments in the style “I would never endorse X, but I don’t understand why others don’t do X”?
I can do that easily without LLM assistance if you like. But if you want your community to be exclusively endorsers of violence against enemies of a chosen tribe, then you should ban me so you can keep your little tribe of Ted Kaczynski fanboys.
This is one of those cases where the word "but" negates everything that precedes it.
If you think we haven't been moderating the type of posts you're talking about, you haven't been tracking HN moderation lately*—which is fine, why should/would you? But in that case you shouldn't be taking snarky swipes at the mods based on galactically mistaken assumptions.
More importantly, you shouldn't be pointing fingers at others instead of taking responsibility for your own bad behavior. Even if you were right in what you said, it wouldn't justify your breaking the rules. Moreover you have a longstanding pattern of doing this and we've been cutting you slack for years.
Okay, admittedly when I read these things I lose my mind and become a viral host for the nonsense because I feel the need to retaliate against what is clearly some kind of Blue Tribe mobbery. Clearly it’s a mistaken belief that you allow targeted mob-forming on your platform. Actually you’re just drowning under the load. Fine. What I can edit out I shall and I’ll try to keep in mind that you’re trying and failing, and doing this is just participating in the crap.
I’ll follow your comments for a mod log to see and I’ll refrain.
I do think it would justify breaking any rules that allow targeted mob-forming but since that’s not happening I’m happy to stand off.
I think your fun cherry fact is pretty inaccurate. If you're referring to USDA Marketing Order #930, it's basically about setting sales limits in bumper crop years to avoid a situation where so many cherries hit the market that farmers lose money simply by harvesting them. They're free to donate the cherries etc. but again, they would be essentially wasting their own money by putting in the time and effort to harvest them beyond the amount they're allowed to sell.
This is for good reason though. You want to overproduce significantly in ordinary times so that if there is a big negative shock you will still be able to produce enough to feed everyone merely by not destroying the excess anymore.
But in a pure market that would mean that during overproduction times, prices should be low. Which they artificially aren't through industry price fixing.
The result that free markets are Prato optimal, though, requires conditions like low barriers to entry, perfect information, and low cost transactions… none of which seem very well met in the case of agriculture.
That's a nice bit of trivia but it doesn't really affect the comment you're replying to. It's still food, full of flavor and calories, and able to be used by a home cook (by making a pie).
If you researched this regulation even a little, you'd see the crops are rarely destroyed. They are far more often exported, diverted to secondary markets, donated, or carried-over into next-season's stock.
It's interesting to me how people are quick to comment about things they know nothing about...
> It's still food, full of flavor and calories
Tart cherries have about 1-2 calories per cherry, and do not taste good without a lot of sugar. That's why they are used in commercial processing, not generally sold as a fruit in grocery stores.
Coming back later, I realized earlier I looked up the calories but I didn't compare them to anything else. So while tart cherries "only" have 50 calories per 100g, sweet cherries are up around 60, not very different. An apple also has about 50-60 per 100g. So does an orange.
Fruit isn't super dense in calories to begin with because it has so much water, but it's still a meaningful amount, and tart cherries are pretty standard among fruit.
So we're moving goalposts? Where did I say people in need don't need any fruit?
People in need don't need single/one calorie tart cherries that are rarely eaten on their own. Consuming tart cherries typically involves processing that is more costly in terms of ingredients and time than simply using the pre-processed versions. Tart cherries are sometimes donated and are rarely destroyed.
Which argument will you come up with next?
You've bounced all over the place in this thread. Just let it rest...
> So we're moving goalposts? Where did I say people in need don't need any fruit?
You gave calories as a reason people don't need this fruit.
But that logic would apply to almost fruit.
So I said it would be bad to say people in need don't need fruit, while pointing out that contrast. I'm not accusing you of thinking that, I'm accusing you of using flawed logic.
> People in need don't need single/one calorie tart cherries
There's plenty of calories in a reasonable serving, and again that argument would apply to almost any fruit. It's like complaining about a single blueberry having too few calories.
> are rarely eaten on their own. Consuming tart cherries typically involves processing that is more costly in terms of ingredients and time than simply using the pre-processed versions.
They can cook with them. Lots of things are rarely eaten on their own and need to be processed, costing more ingredients and time than the pre-processed form. This includes flour!
> Tart cherries are sometimes donated and are rarely destroyed.
This is true and has nothing to do with my point.
> Which argument will you come up with next?
If you bring up a new reason to imply that donating tart cherries is unreasonable (even though it does happen!), I might disagree with that reason. Otherwise I have had one single argument and it hasn't changed: Donating tart cherries is a good idea.
I don't know why you're so fixated on whether people eat something directly. That doesn't affect what all2 was saying or what voxl was saying or what anyone else has been saying, but you keep acting like it does.
So you understood the crop we're discussing is rarely destroyed - and more often donated, diverted to secondary markets (ie. sold in grocery stores), or exported - yet still felt compelled to say a home cook could use them?
What was even the point of your snarky comment then?
> So you understood the crop we're discussing is rarely destroyed - and more often donated, diverted to secondary markets (ie. sold in grocery stores), or exported - yet still felt compelled to say a home cook could use them?
In the context of someone talking about home cooks using them, and you acting like "People do not eat tart cherries directly." is a counterargument, yes I felt compelled to correct that.
The incorrect thing you were implying had nothing to do with how often they're actually destroyed. So why would that stop me?
People do not eat tart cherries directly. The overwhelming majority of people will never process them into something edible either.
"People in need" are not going to spend time and money processing tart cherries into juice concentrate or pie filling... especially when a can of either is cheaper than the raw ingredients to make your own.
Your point is ridiculous, absurd and pedantic beyond any reasonable purpose.
Most of what you are saying is correct, but I feel the need to respond to your far too many repeated assertions that "People do not eat tart cherries directly": Except for when they do!
I grow several varieties of sour cherries in my yard, and frequently use them whole and without further processing. Usually I use them in a recipe like this: https://en.wikipedia.org/wiki/Clafoutis. Sometimes I pit them first, sometimes I don't. Sometimes I'll even happily snack on them raw.
No, like most small fruit you aren't going eat them because you are desperate for calories. But they actually aren't any harder to prepare or use than lots of other tasty things that people traditionally grow.
reply