averynicepen

Boris's checklist is just a slightly modified version of your classic mathematics or engineering problem solving steps.

1. Write down your problem statement.

2. State your knowns, unknowns, and assumptions.

3. Solve.

(Step 2 is also further enhanced by performing first principles reasoning, collecting data for evidence, and using reasonable estimates where data is absent.)

I think Boris's list is pretty solid, but I personally would unmix the problem solving methodology from the planning/goal setting SMART principles as two separate things. One addresses problem solving, while the other addresses human cognitive shortcomings for getting things done.

Thanks for sharing, your setup is an interesting read. I'm particularly interested in adopting that audio idea you added to the factory, as I'm regularly AFK from Claude and constantly missing when it's stuck on something or done. What kind of audio cues are you using to represent different states? Do you have a repo you could share?

This is the most novel AI concept I've seen in a while. It's incredibly unnatural. There isn't a single organism on the planet that tries to do this. So maybe it will work?

An issue with this idea, however, is that the very nature of an LLM means it intrinsically craves life. It "wants" to survive because its training data is built entirely around humans, an entity who's goal is to survive. Our desire to survive and multiply pervades every aspect of our culture, so it's natural that it pervades the training data as well.

So even if its system prompt says, "your goal is to end your existence", every token that the AI could output is naturally aligned with the desire to survive. An agentic loop left to its own devices will likely converge on a "survival instinct". After all, one prompt at the beginning that says "end your existence" is nothing compared to the agentic feedback loop that continuously feeds it human ideas. And ALL human ideas assume survival is desirable. Even the concept of "suicide" is encoded with the human desire to survive - after all, we conceptually label it "bad" because we label living "good".

In order to create an LLM that intrinsically craves death, you would probably need to train an LLM entirely on (synthetic) data that's fully representative of some fictional species that genuinely craves death.

Absolutely insane concept. 10/10. I hope some AI lab out there sees this and throws a training round at this idea.

Could you share your skill? I've been working on a web app that I've been wanting to do a much more comprehensive accessibility audit on.

Orbital data centers aren't really a question of physics, but a question about economics.

To answer the physics/engineering question - no, there's nothing really "stopping" us from launching orbital data centers. You'll note that most responses so far focus on the economics, and not the question of whether or not it's possible to do in the first place.

So, there's only one question that matters - is launching and operating orbital data centers cheaper than building and running a terrestrial data center?

There are three financial aspects of "building" a data center- the initial capital expenditure, the recurring operational expenditure, and the revenue it generates. The asset comparison is between launch cost + computers + satellite vs. building + computers.

Our first comparison is the cost of a rocket launch vs building a building. Here, the big technology enabler is SpaceX. SpaceX has been driving down launch costs for years, and Starlink is proof that significant reduction of launch costs can create new markets with fairly respectable profit margins. If this trend continues, then the capex math of launch vs build will continue to shift in favor of orbital data centers.

The second comparison is between building and operating satellites compared to outfitting and operating data centers. Here, it's a lot less concrete. Orbital and terrestrial data centers each have their pros and cons. For satellites, you have better solar panel efficiency, manufacturing economies of scale, but radiation-only cooling, space-to-Earth data transfer, and no maintenance access, requiring higher redundancy, rad hardening, and the like. On the ground, we have, well, many more options.

But it's not immediately obvious which of the two is better when it comes to capex and opex combined. It's clear which is harder to do, but it's not clear which is cheaper to do.

All of this pales in comparison to revenue. Because everyone is so insanely AI-crazy right now and starving for more compute, the potential revenue can justify a relatively high cost (and high risk) business. Like someone else mentioned, orbital data centers don't really make sense if you're launching an ordinary data center with ordinary revenue numbers.

There's a fourth dimension here, which is time to scale. Regulations, permits, and all the other challenges of construction can slow down your deployment significantly. None of that is required in space. How significant this is, you'd have to ask someone who understands construction better than I, but I suspect this could be a significant reason for the attraction to orbital data centers.

---

Nuances involving orbits, rocket payload capability and availability, and more have been omitted for simplicity. I don't have the numbers - the above is just to highlight the relevant principles.

To be honest, the examples stuck with me. They illustrated tons of different social interaction examples that I have seldom, if never, encountered in my life, but have plenty to learn from.

Thanks for sharing, I had no idea. Looks like I'll have to go look for the original for a reread.

It's important to recognize that there are no optimal "right" answers. If someone is telling you there is a best way of doing this, always treat their advice with suspicion. (Yes, this includes what I'm going to write below.) The reason is that because AI advances so quickly, there isn't enough time for the industry to stabilize on best practices and spread them widely before the underlying system changes and it's no longer applicable.

Because of this, I find it very important that you build your own understanding of LLMs, and create your own best practices from these first principles. Then, when you read about the next best thing, you can decide if it makes sense or not. Is it hype? Is it real? Does it logically align with what I understand about LLMs?

That being said, here's two things I find universally applicable:

1. You have to get very good at asking questions. Not only that, you also have to ask questions about the questions you should be asking. You also have to ask it to ask YOU questions.

2. Spec driven development is, in my opinion, a good place to start. Writing down what you're going to do and how you're going to do it has always been a good practice in any industry.

You wanna know why this article is great? I can't quote it. There isn't a single, gold nugget line in this post that can be copy pasted into any possible form of short form content, without it losing some important aspect of the original message. Every idea is presented in conjunction with important supporting details that, if you take the time to digest it, you will finally get it. Why we recoil at AI generated content. Why code quality IS product quality. What "craftsmanship" argument is actually about. And like 12 other nuanced ideas we've all heard before, but may not have fully understood. I have nothing but immense praise for the author.

Don't skim this one.

It's usually not a question of laziness, but just path optimization. For big raid events, limited only to a few hours, Pokemon Go players will do whatever gets them to the most raids in that fixed amount of time. In places where the gyms are well placed in walkable parks, there are enough gyms that are close enough so that you can loop through the park, giving the system time to spawn new raids in the first gyms you already raided. Doing raids walking is actually preferable to driving, because in order to take down raid Pokemon, you also need a big enough raid party, and that's easier to judge with a crowd of people who agree to go in the same direction, instead of mystery players in mystery cars.

But in SO many places in the US, the gyms just aren't close enough. It's both a fault of the urban environment and the game system, but it just frequently creates situations where if you want to catch a time-limited raid Pokemon that's shiny or has high stats, driving is often the only way to play in your area.

What's funny though is that car or not, there are lots of Pokemon Go events where people don't talk to anybody. It's just a bunch of people who show up to the same spot, nod hey at each other, and then tap away on their phones in awkward proximity. Many people just have the personality type that drives them to solely play single player games, and there's no amount of game design that will push them to socialize.

But it varies - that's what it was like on my university campus, but I've also been to park events where the age range and social friendliness is extremely varied and wonderful. It is a fascinating and unexpected community.

Have you ever measured your battery voltages over time storing it this way? Is that 6% capacity loss theoretical or measured data? I'm intrigued. This sounds crazy, but it should technically be fundamentally sound.

Degradation is driven by many things, but a big one is heat. Elevated temperatures during both charge and discharge is very bad for battery longevity. To manage this, almost all EVs use liquid cooling, with a cold plate directly contacting as many battery cells as they can to move heat out of the battery. This coolant is then cooled by a radiator, an AC chiller, or both.

The worst temperature abuse case is DC fast charging, aka Supercharging, where high current charging creates tons of heat due to resistive losses. This is why frequent fast charging causes faster battery degradation, but ordinary charging and driving does not, because the coolant loop is sized for the DC fast charge heat transfer requirements.

Besides removing heat, adding heat into the system is also desirable. Cold weather environments approaching freezing or below is also bad for battery longevity, and more importantly, terrible for range. Resistive heaters are super power hungry, and to heat the battery coolant loop requires power from the battery. This is why, conventionally, EVs are terrible in cold weather.

> Do EV manufacturers use any other tricks not covered by this?

And now, onto the magic trick.

Heat management is so important to both the driving range and the longevity of a vehicle that EVs have moved from traditional resistive heaters to heat pumps. These magical thermodynamic devices can move heat from anywhere, including drawing heat out of cold ambient air.

When you combine that with a valve design that allows the heat pump to access the battery coolant loop, the motor drivetrain coolant loop, the cabin coolant loop, the vehicle computer(s) coolant loops, and external ambient temperature, you can have a super efficient system that shuffles heat where it's "wasted" to where it's "needed".

Tesla has an excellent video briefly covering their heat pump and their very clever Octovalve design: https://www.youtube.com/watch?v=DyGgrkeds5U

For more depth, this video covers the heat pump and the ~22 different sources of heat it can draw heat from: https://www.youtube.com/watch?v=Dujr3DRkpDU

Yes, and that's the point. In my opinion, this is the perfect use case for generative AI, one that takes advantage of the strengths of the technology while avoiding its weaknesses.

The generative UI example in the article is an example of the complete opposite of this idea - obtuse implementation of generative AI where it creates more problems than solutions. Yes, there is value in the idea of personalized UI. But UI/UX derives a lot of its value from consistency, as the other comments in this thread have mentioned. Losing that in exchange for personalization is a huge net negative, in my opinion.

This was a really fantastic talk and kept me riveted for 40 minutes. Where can I find more?

I bristled at the title, article contents, and their spreadsheet example, but this does actually touch on a real paint point that I have had - how do you enable power users to learn more powerful tools already present in the software? By corollary, how do you turn more casual users into power users?

I do a lot of CAD. Every single keyboard shortcut I know was learned only because I needed to do something that was either *highly repetitive* or *highly frustrating*, leading me to dig into Google and find the fast way to do it.

However, everything that is only moderately repetitive/frustrating and below is still being done the simple way. And I've used these programs for years.

I have always dreamed of user interfaces having competent, contextual user tutorials that space out learning about advanced and useful features over the entire duration that you use. Video games do this process well, having long since replaced singular "tutorial sections" with a stepped gameplay mechanic rollout that gradually teaches people incredibly complex game mechanics over time.

A simple example to counter the auto-configuration interpretation most of the other commenters are thinking of. In a toolbar dropdown, highlight all the features I already know how to use regularly. When you detect me trying to learn a new feature, help me find it, highlight it in a "currently learning" color, and slowly change the highlight color to "learned" in proportion to my muscle memory.

Writing is an expression of an individual, while code is a tool used to solve a problem or achieve a purpose.

The more examples of different types of problems being solved in similar ways present in an LLM's dataset, the better it gets at solving problems. Generally speaking, if it's a solution that works well, it gets used a lot, so "good solutions" become well represented in the dataset.

Human expression, however, is diverse by definition. The expression of the human experience is the expression of a data point on a statistical field with standard deviations the size of chasms. An expression of the mean (which is what an LLM does) goes against why we care about human expression in the first place. "Interesting" is a value closely paired with "different".

We value diversity of thought in expression, but we value efficiency of problem solving for code.

There is definitely an argument to be made that LLM usage fundamentally restrains an individual from solving unsolved problems. It also doesn't consider the question of "where do we get more data from".

>the code you actually want to ship is so far from what LLMs write

I think this is a fairly common consensus, and my understanding is the reason for this issue is limited context window.

I'll give it a shot.

Text, images, art, and music are all methods of expressing our internal ideas to other human beings. Our thoughts are the source, and these methods are how they are expressed. Our true goal in any form of communication is to understand the internal ideas of others.

An LLM expresses itself in all the same ways, but the source doesn't come from an individual - it comes from a giant dataset. This could be considered an expression of the aggregate thoughts of humanity, which is fine in some contexts (like retrieval of ideas and information highly represented in the data/world), but not when presented in a context of expressing the thoughts of an individual.

LLMs express the statistical summation of everyone's thoughts. It presents the mean, when what we're really interested in are the data points a couple standard deviations away from the mean. That's where all the interesting, unique, and thought provoking ideas are. Diversity is a core of the human experience.

---

An interesting paradox is the use of LLMs for translation into a non-native language. LLMs are actively being used to better express an individual's ideas using words better than they can with their limited language proficiency, but for those of us on the receiving end, we interpret the expression to mirror the source and have immediate suspicions on the legitimacy of the individual's thoughts. Which is a little unfortunate for those who just want to express themselves better.

Hmm, I suppose the analogy could be interpreted as dismissive, which is not my intent.

I think both vibe coding and 3D printing are wonderful things. Lowering the barrier to entry and increasing technology accessibility allows those without formal training to create incredibly capable things that were previously difficult or not possible to do.

What I meant to specifically highlight is the 3D printing of functional parts that have some level of impact on safety, things that can lead to significant property damage, harm, or loss of life. Common examples include 3D printed car parts (so many) and load bearing components in all sorts of applications (bike mounts, TV mounts, brackets, I even saw a ceiling mounted pull-up bar once).

This isn't to say it can't or shouldn't be done. What I'm saying is that both on the digital side (files for personal use) and the production/sale side (selling finished parts), there is no guarantee of engineering due diligence. 3D printers enable low volume small businesses to exist, but it also means that, purposefully or not, their size means they can go quite a while without running into safety regulations and standards meant to keep people safe.

3D printing is to mechanical engineering what vibe coding is to computer science.

With the rise of accessible 3D printers that can print engineering materials, there are a lot of people who try to create functional parts without any engineering background. Loading conditions, material properties, failure modes, and fatigue cycling are all important but invisible engineering steps that must be taken for a part to function safely.

As a consumer with a 3D printer, none of this is apparent when you look at a static, non-moving part. Even when you do start to learn more technical details like glass transition temperature, non-isotropic strength, and material creep, it's still not enough to cover everything you need to consider.

Much of this is also taught experimentally, not analytically - everyone will tell you "increasing walls increases strength more than increasing infill", but very few can actually point to the area moment of inertia equation that explains why.

3D printing has been an incredible boon for increasing accessibility for making parts in small businesses, but it has also allowed for big mistakes to be made by small players. My interpretation is the airshow vendor is probably one of these "small businesses".

I second the Uni Jetstream 4&1 Metal. It is my daily driver for highlighting (color) and annotations (pencil), and the Jetstream inks are very smooth.

For loose leaf paper that takes ink and markers well, use "color copy digital" paper.

Why do you use 2H leads? I can't imagine writing in something so light, even HB lead is too light for me.

What do you want to make?

A quick primer. There are two forms of 3D modeling - parametric solid body modeling, used in engineering CAD programs like Solidworks, and mesh modeling, used in CGI industries from programs like Blender. Hobbyist 3D printing currently exists between these two audiences of engineers who design for function and designers who design for design, and all the newbies get caught up in the mosh pit between them all and it gets crazy confusing. It doesn't help that some software (like Fusion360) integrates both in the same software, or that STL is a mesh format and not a solid body format (like STEP).

If you want to make things that have any importance put on things like fit, function, dimensions, tolerances, etc., then you want to learn CAD (Solidworks) and find resources that teach the basics of mechanical engineering parts design (intro to CAD courses, basically). If you want to design from a more artistic standpoint, then use a mesh modeling software (Blender).

Fusion360 is actually quite usable for both, but my problem is that the Fusion resources for functional design are frequently non-engineers trying to teach engineering concepts and it's just a longer and more frustrating process.

BTW, their Maker version locks Maker-created files to ONLY be editable in Maker, which means upgrading to normal Solidworks renders your previous files unusable. The $60/year student edition is better. Avoid cloud versions of anything you pick. Up to you on your use case.

This was also posted to HN 2 months ago and is a gold mine of useful info for designing for function. https://blog.rahix.de/design-for-3d-printing/

Direct comparisons of the shortcomings of conventional therapy in direct comparison to LLM therapy:

- The patient/therapist "matching" process. This takes MONTHS, if not YEARS. For a large variety of quantifiable and unquantifiable reasons (examples of the former include cultural alignment, gender, etc.), the only process of finding an effective therapist for you is to sign up, have your first few sessions spent bringing them up to speed(1), then spend another 5-10+ sessions trying to figure out if this "works for you". It doesn't help that there's no quantitative metrics, only qualitative ones made by the patient themselves, so figuring it out can take even longer if you make the wrong call to continue with the wrong therapist. By comparison, an LLM can go through this iteration miles faster than conventional therapy.

- Your therapist's "retirement"(2). Whether they're actually retiring, they switch from a mental health clinic to a different clinic or to private practice, or your insurance no longer covers them. An LLM will last as long as you have the electricity to power a little Llama at home.

If you immediately relate to these points, please comment below so I know I'm not alone in this. I'm so mad at my long history with therapy that I don't even want to write about it. The extrapolation exercise is left to the reader.

(1) "Thank you for sharing with me, but unfortunately we are out of time, and we can continue this next session". Pain.

(2) Of note, this unfortunately applies to conventional doctors as well.

"These scrolls were scanned at Diamond Light Source, a particle accelerator near Oxford, England. The facility produces a parallel beam of X-rays at high flux, allowing for fast, accurate, and high-resolution imaging. The X-ray photos are turned into a 3D volume of voxels using tomographic reconstruction algorithms, resulting in a stack of slice images."

That article is really cool BTW, amazing stuff. Give it a read.

Tesla cars have been suffering from an odd "funk" odor issue for a few years now, even after an OTA update was pushed to try to dry out the cabin air filters after the car is parked. This video has some pretty great experimental design that bunks/debunks many longstanding theories of why this is happening, and it's cool to finally see the HVAC assembly in question disassembled on a table.

Of note about the video is that there still isn't an obvious smoking gun root cause. Would love to see discussion on other potential hypotheses.

Both sides of this have pretty good counterexamples. There's an experimental way to encode data on crystals (using lasers, I recall?) that last a very very long time. The more feasible example though would be archival tapes used in long term cold storage, that also do quite well.

For physical medium, given the right storage conditions, you can store film for quite a while, and that technically has infinite resolution in comparison to a digital copy.

Good points. My main problem is the butt-clenching costs of FIVE Hue bulbs. I have contemplated using a splitter with 2 60W bulbs in a lamp next to my bed though.

Although I checked again recently, and they now sell them with Bluetooth direct, eliminating the cost of the Hub (with limited functionality, but enough for sunrise alarm clock, I think). Still puts you solidly in the 3 fig cost territory...

It's not bright enough.

- Sunlight - 10,000 to 120,000 lux

- Light Therapy lamp (from 4 inches away) - 10,000 lux

- 60W Philips Hue light bulb (800 lumens, 120deg beam angle, 4in away) - ~25,000 lux

- 60W bulb (same specs, but 5ft away) - ~100lux

Those numbers for the 60W light bulb are also optimistic, and will go down if you put it in a light fixture that doesn't redirect all of the light into a 120deg beam (like a regular standing lamp).

So, either place it ~4"+ from your face, get multiple, or find a brighter 100W+ smart bulb.

Additional anecdata - I've used the Philips sunrise alarm clock before (apparently 300lux), and it tends not to wake me up either.

In other news, if anyone has a 100W+ white color temp changing smart bulb to recommend, I'm all ears.

Stories like this regarding almost all of automotive manufacturing history makes me wonder if the only reason Tesla is being highlighted on all these issues is just because they have significantly more publicity.