I've traveled to Hanoi and absolutely loved the phở there. One perk of living in Berlin is the exceptional Phở we get here. Thanks to the large Vietnamese community we have great phở, Bahn mí (esp at the Dong Xuan center). The phở we get here is top tier too, as confirmed by my Vietnamese friends.
I hereby certify that these are indeed the most perfect and precise svg depictions of pelican riding a bicycle, also known among biology scholars as pelycles
> If a model finally comes out that produces an excellent SVG of a pelican riding a bicycle you can bet I’m going to test it on all manner of creatures riding all sorts of transportation devices.
This relies on the false premise that, if they would include it in their training dataset, it would be perfect. All they need to do is be good enough and better than the other, not perfect.
I'm not sure if we can have a "perfect" Pelican riding a bicycle. Like, I could probably commission a highly experienced artist to draw one and I don't think it would be perfect. The legs would probably have to be too long, or pedals oddly placed, or handles strange, or wings with hands.
Based on the one Simon commented though, I'd say we're in decent territory to try the latter part of his hypothesis.
> The legs would probably have to be too long, or pedals oddly placed, or handles strange, or wings with hands.
In all seriousness, that's what makes it an interesting test: it's asking for something technically impossible, that requires artistic license to make coherent.
Making specific choices on where to bend reality (and where not to) is a big chunk of visual art.
I mean the prompt was succinct and clear, as always - and it still decided to hallucinate multiple features (animation + controls) beyond the prompt.
It'd also like to point out that to date no drawing was actually good from an actual quality perspective (as in comparative to what a decent designer would throw together)
Theyre always only "good" from the perspective of it being a one shot low effort prompt. Very little content for training purposes.
The way I’ve come to think of LLM is that what the produce in a single reply even with thinking turned up, is akin to what you’d do in a single short session of work.
And so if you ask it to do something big it will do a very surface level implementation. But if you have it iterate many times, or give it small pieces each time, you’ll end up with something closer to what a human would do.
I imagine the pelican test but done in a harness that has the agents iterate 10+ times would be closer to what you’d expect, especially if a visual model was critiquing each time.
Yeah, this is how I use AI. Instead of a single session one-shot, it's usually limited to single targeted edits, and then I steer it on each step. Takes longer but the output is actually what I want.
What does good even mean… I have no idea what a good “pelican on a bike” should look like. It’s a fun prompt because there is no good answers… at least so I thought.
I tried GLM5.1 last week after reading about it here. It was slow as molasses for routine tasks and I had to switch back to Claude.
It also ran out of 5H credit limit faster than Claude.
If you view the "thinking" traces you can see why; it will go back and forth on potential solutions, writing full implementations in the thinking block then debating them, constantly circling back to points it raised earlier, and starting every other paragraph with "Actually…" or "But wait!"
I was watching Qwen3.6-35B-A3B (locally) doing the same dance yesterday. It eventually finished and had a reasonable answer, but it sure went back and forth on a bunch of things I had explicitly said not to do before coming to a conclusion. At least said conclusion was not any of the things I'd said not to do.
That is essentially what the reasoning reinforcement training does. It is getting the model to say things that are more likely to result in the correct final answer. Everything it does in between doesn't necessarily need to be valid argument to produce the answer. You can think of it as filling the context with whatever is needed to make the right answer come out next. Valid arguments obviously help. but so might expressions of incorrect things that are not obviously untrue to the model until it sees them written out. The What's The Magic Word paper shows how far that could go. If the policy model managed to learn enough magic words it would be theoretically possible to end up with an LLM that spouts utter gibberish until delivering the correct answer seemingly out of the blue.
That's pretty cool, thanks for the extra context! (pardon the... not even pun I guess)
Also, thanks for pointing me at that specific paper; I spend a lot more of my life closer to classical control theory than ML theory so it's always neat to see the intersection of them. My unsubstantiated hypothesis is that controls & ML are going to start getting looked at more holistically, and not in the way I normal see it (which is "why worry about classical control theory, just solve the problem with RL"). Control theory is largely about steering dynamic systems along stable trajectories through state space... which is largely what iterative "fill in the next word" LLM models are doing. The intersection, I hope, will be interesting and add significant efficiency.
Spandau occupies a commanding position on the wide confluence of Spree and Havel river. It is part of the wider river networks and one can easily navigate to Elbe and Danube. The coin might have ended up here via trade.
This is so important. Kerala in particular had a treasure trove of varieties, some well suited to low rainfall, resistant to local pests. I am sure other states had/has such diversity too, I am just not knowledgeable enough.
These genotypes are being lost to industrial mono-cropping. The government is doing nothing about it.
> These genotypes are being lost to industrial mono-cropping. The government is doing nothing about it.
This is happening worldwide and is one of the tragedies of modernity. Mexico for instance has tons of regional varieties of peppers that don't grow anywhere else except for in a very specific micro climate and they're disappearing in large part because of cheap imports that makes farming them unprofitable.
I grew up in similar environment in rural central India and I (half)jokingly say farming was my first real job. We rarely planted rice, but I have vivid memory of helping my father plant the rice saplings in the muddy puddles in my farm.
I am always skeptical of urban people wanting to move back to little villages to do farming. Farming is a back-breaking and a tough job. You are exposed to all the vagaries of nature. The market forces are also not always in your favour. It is another version of "quit-job-and-open-a-coffee-shop" fallacy.
Well, those who quit their jobs and open coffee shops almost certainly make a bad choice for themselves economically and work/ life balance wise... But they do wonderful things for their community, and - a questionable benefit to society but a huge benefit to some - real estate prices. People love these places. They capture a tiny fraction of the value they create, if we look at it in cold terms.
That can't really be said for downscaling rice farmers, can it? I mean, at best maybe the other rice farmers enjoy having them around.
I meant it as a pipe-dream that people jump into without knowing the hidden asymmetries. Farming of any kind is hard and learning that specific skillset is necessary to succeed.
For most folks it's just an add on. I have grandparents in Europe that have a garden where they grew potatoes and about 50 other things I'm not gonna list. They make jams, pickled things, and various other preserves. It's something to do and kept them sharp until they hit their late 80s.
Agreed. There's a world of difference between 'farming' for personal to small scale production as not quite a recreation but also not quite a job, and farming a low margin staple at high volume as your primary and sole means of earning money.
And I think when most people speak of the dream of returning to rural society to e.g. farm, they're speaking very much of the former rather than the latter.
That has been my experience as well, having immigrated from Eastern Europe to an enclave in the US. We know at least a dozen families (including our own) with 2-10 acre homesteads and all of them had previous experience with gardens and dachas in the Soviet Union that they used to grow supplemental produce, so no one came into the deal with delusions of making any profit. Everyone gives away the excess to neighbors of which there is usually a lot because yields are high on hand tended trees (and dutch bucket hydro).
The single biggest reason these farms exist is because American retail produce is mostly garbage. It’s so economically micro-optimized that all flavour has been wrung out of it. The only way many of us immigrants can get back the flavors of our childhoods is by growing the fruits and vegetables ourselves, if only to have control over the varieties, the vast majority of which are not sold in stores (>95%). That nostalgia is what pays the margin.
We're not neighbors unfortunately because we're spread out all over Southern California. By "enclave" I mean the area between West Hollywood and Arcadia, where many Eastern Europeans immigrated during the post-Soviet brain drain, not a dense conglomerate like San Gabriel.
BTW you do NOT want ten acres. That is a back breaking amount of work and even with modern technology you'll struggle to cope (it's not enough to afford most heavy equipment, but too much to do manually). You want an acre or two where you have enough space to plant trees. It takes a few years from nursery to fruiting, but they are far lower maintenance.
You don't really need 10 acres. My grandparents made do with 1/4 of an acre and would have yields of 350-500 lbs of potatoes per season. That's so much that they would give it away. I have fruit trees that require almost no effort to maintain once established. My neighbors give me oranges that fall to the ground and rot otherwise. It's not all or nothing. You can have a basil plant in an apartment.
Lots of places have community gardens. Hell, I go to one in the middle of NYC, a rooftop garden run by a friend. We even grow our own wheat for bread making.
> You have a right to perform your prescribed duty, but you are not entitled to the fruits of action. Never consider yourself the cause of the results of your activities, and never be attached to not doing your duty.
I have a hard time interpreting that as what geohotz is saying. If anything it seems to promote rent seekers by telling you - stick to your lane and don't complain. I.e. the caste system
> If anything it seems to promote rent seekers by telling you - stick to your lane and don't complain. I.e. the caste system
I was wondering if that would come up and HN delivers without fail. Anyway, you are free to interpret it as you see fit.
The guidance was for someone who was struggling with a moral dillema on facing relatives in war and undecided over action. It is not a diktat to work or provide unquestion labor.
For anyone who understood the whole story and backdrop of the situation, a reasonable interpretation is
- you are responsible for your actions but you cannot control the consequences of your actions due to many factors.
- When you detach yourselves from results, you can do your job without anxiety.
- do not let the fear over results be an excuse for inaction.
Give it a read and decide for yourselves if you are not convinced. Even without the teachings part, the whole story of Gita is actually an epic story/novel with some strong and conflicted characters with elaborate back stories.
The point is worrying about results creates anxiety that takes away your focus from doing the job which results is lower performance -> subpar results - it is a recursive spiral. Also results are not "wholly" under our control. Hence avoid worrying about the results "while" doing it.
The final result of any action depends on a lot of external factors for e.g. one might develop a product, but whether it succeeds or not depends on other players, economy, luck and host of other factors that one does not control or know. This feeds into the first point about not worrying about the results.
The idea is to not fret over results and give your best with dedication. E.g. An athlete shouldn't be worrying about their results while playing and should focus on the play.
> How can you be responsible for your actions but not the consequences of your actions
I only mentioned control of consequences not responsibility. It doesn't mean you are absolved of responsibility from consequences of your actions.
Take driving. You can only control your actions but you cannot control what happens on the road. You are still responsible for your actions.
Should the fear of the unknown on the road stop you from driving? Absolutely not.
Yeah, I think GP is right and this all feels like word play to me.
Let's say you're making your ideal wood chair for your living room. You're "fretting" over getting the arms, legs perfectly aligned, surfaces smooth, identical chamfers, etc. But fretting is bad and makes it hard to get anything done so you stop. Then you end up with a roughshod chair.
You shouldn't "fret" too much, but you also shouldn't "fret" too little. You should "fret" exactly the right amount which is... tautologically true, and not particularly insightful. Giving your best is committing yourself to fretting.
(I'm not sure if you understood the Gita reference, if you did you can ignore my comment.)
A better analogy isn't so much a chair made by yourself, but say you're making a chair for a competition. The only thing you can control is your actions. What's out of your control is whether or not some other guy who's an absolute veteran is also planning to compete (making you likely to lose), or if almost no one else joined the contest (much higher chance of winning.)
Another example is in the workplace.
You control what efforts you put in.
But your promotion (or even firing!) depends on a ton of other things. If the company's doing poorly, they may not promote you even if you did a great job. Conversely, if the company desperately needs a new senior person and you're the closest fit, even if you just did a mediocre job, you may get selected.
What's common in both cases is - doing a better job is more likely to be better for you. But there are no guarantees you'll win/get promoted, so don't be attached to that outcome, but only your efforts/labour.
> You have a right to perform your prescribed duty, but you are not entitled to the fruits of action.
> stick to your lane and don't complain. I.e. the caste system
That verse is quite famous and the general interpretation as I understand is this.
You have control on your actions but not on its results. The results depend not only on your actions but on many other factors outside of your control.
Now, one can interpret that it is instruction to "stay in your lane", but I have not seen that interpretation so far in my life in India.
you realize that the caste system that currently exists is completely different from what it was conceptualized as right? you most certainly want to read the conversation between yudhistra and nahusha that talks about caste https://vedabase.io/en/library/mbk/1/30/
I don't think we actually know who conceptualized the caste system. Even Manusmriti seems to be not as old as we thought before.
However, even in Mahabharatha there are examples of Karna and Ekalavya who despite having qualities (as Yudhishtira claims) of Kshathriya, they were rejected by the society as being lesser.
It is possible that caste system is an extension and crystallization of nepotism. Typically professions and trade secrets are handed down the families and it is conceivable this was codified at some point far in the past.
To claim that caste system has a more philosophical foundation would be a bit of a stretch in my point of view, especially when it has been throughout the history being used suppress.
You need to understand the context. The quote in Gita was to motivate the best warrior of the time at the battlefront facing opponents who were mainly his cousins and uncles.
In that context the quote is about performing the duties you were born to do without overthinking the consequences.
the context makes it even worse. its a strange kind of tribalism that is being promoted here. "do what you are asked to without understanding the real consequences". btw war is actual zero sum usually.
It’s a tricky philosophy to put into practice. I have oscillated between this approach (“owning” the effort) and “owning” the outcome. I have found that taking ownership of the outcome leads to better results because I have a personal stake in the outcome and I tend to think through the problem more deeply, but I am almost always left feeling more stressed and “empty” when the work is finished. When I focus on doing the best I can and let go of the outcome, the end result is almost always subpar which leaves me feeling frustrated, because I know it could have been better had I taken on more responsibility.
You’re right, it was 2014. I was there on HN when docker was announced by shykes. It was a godsend because I was getting bummed by the alternatives like LXC, juju charms or vagrant.
I still prefer LXC to docker. Improving libvirt and making virtualization a first-class OS feature with a library interface - vs. relying on an external tool and company interested in monetization - was and is the right approach.
reply