HNHacker News
TopNewBestAskShowJobs

jjk166

8,522 karma · joined April 26, 2017

submissionscomments
jjk166··on CEOs who think AI replaces their employees are just bad CEOs
Because it isn't. Aligning incentives is really easy.

"If you do something that causes a productivity gain worth X, I will give you some reward with a value Y where Y is less than X but greater than the cost Z of the effort you needed to put in to generate the productivity gain. If the cost Z would be equal to or greater than X then don't do it."

Managers make their work immensely harder on themselves by unnecessarily adding the constraint that they can't get people to do things by paying them fair amounts to do them. Now certainly there are some highly skilled managers out there who can still succeed despite this handicap, and if that earns them a fat paycheck then good for them. But if you don't have those skills you don't get to excuse your failures with an inefficiency you created for yourself.

jjk166··on Unified Controllable and Faithful Text-to-CAD Generation with LLMs
> Download model, done.

Note how that is not a parametric model with the appropriate variable parameters nor does it have the set screw I specified in my prompt. Even if precisely the model you want has been made by somebody else, it still takes time and effort to hunt it down. By the same logic who needs to generate 2D images when stock photos are a thing?

> If you are making something complex enough to warrant you yourself designing it (something unique, custom, new), then CAD on its own is the best tool (right now!).

I make CAD models every day, professionally. Most are not complex at all. Unique, custom, and new don't imply any level of complexity, and none of the above correspond to difficulty to describe. As a simple example from today, I need a funnel with a specific volume and sizing for the inlet and outlets. No one makes it, there is no model I can download, it is unique, custom and new. The modelling itself is not hard but the tricky bit is doing the math to ensure I get the right volume without compromising the geometry. The CAD software solves the equation but I had to spend time and effort to figure out the equation to type in. This is exactly the sort of task that LLMs excel at.

> I have models with thousands of parametric features linked to hundreds of other parts in an assembly. I would need to write a novel to achieve those same features with an LLM.

Did you hit a button to create that whole assembly and form all those links with a single short action, or did you spend hours if not days meticulously making or modifying models for each of those parts then manually link them each together to create those relationships? You put in the time and effort to write a novel, you just wrote it in a different language.

> If what you want to make is simple enough that an LLM would be able to nearly one-shot it, then that model either: 1. Already exists somewhere (like the gear above) or 2. Is simple enough that a couple hours of training videos will give you all the skills you need to make it yourself

You've got it completely backwards. CAD tools are well set up for doing standard work, where LLMs would excel would be where you're departing from the normal to do more complex stuff. Simple example: let's say you want a 10x10 grid of 1/4-20 threaded holes, most any cad software you can do that in under a minute. Now let's say you want 100 1/4-20 threaded holes arranged in a spiral pattern where the distance between each pair increases by 1% as you move outwards. Is that doable in CAD? Most certainly, but it's gonna be a pain in the butt. Does a model exist on the internet? I doubt it, maybe for some generic flat plate but definitely not the part you need this feature on. On the other hand, typing out a prompt to describe it takes virtually no time at all. It's unique, custom, new but much like my funnel it isn't complex, at least not conceptually.

jjk166··on Switzerland wil have a referendum to cap population at 10M
You very clearly did not understand what I wrote, so let's simplify this further.

There is one guy alone on an island. There are two lighthouses on the island. Each lighthouse needs an operator to function. The guy operates one of the lighthouses. How much do wages need to rise to get both lighthouses operating at the same time?

jjk166··on Unified Controllable and Faithful Text-to-CAD Generation with LLMs
Definitely for fine adjustment you're going to still use the manual CAD tools, but you can describe things much faster than you can model them. Good prompt engineering is important here as much as for any other AI use case.

For example let's say I need a parametric model of an involute gear with a 14.5 degree pitch angle and a 0.5 inch hub with an 8-32 set screw where the OD, number of teeth, face length, and shaft diameter are all variable. It takes seconds to write the prompt and then adjust the variables to what I want whereas it would take tens of minutes to actually model even for a highly skilled drafstman.

The time savings get more extreme the more the modelling work is looking up information online. For example if I wanted a model of a lightbulb electrical connector, 95% of the work is just going to be googling. Technically you could have an LLM just tell you all the dimensions you need and then model it yourself, but that's definitely still going to be slower and you have to put in the effort to figure out what each dimension refers to. It makes sense just to cut out the middleman.

Add in the fact that CAD is in fact a difficult tool to use and learn effectively. There is a subset of the population that is very good at picturing and orientating objects in 3D space in their minds, and engineering is mostly limited to these folks. For those whose minds work differently, CAD is extremely tricky to learn. An alternate method of interacting with models that does most of the heavy lifting such that a person only needs to tweak a near complete model could be extremely helpful to many people who would like the benefits of CAD design but have struggled to learn the software. It's no different than AI generated art opening digital art to those who are not good with digital artistry software.

jjk166··on CEOs who think AI replaces their employees are just bad CEOs
Or all 15 people are still employed doing the work of 22.5. Or even more people have been hired now that each person can generate 50% more value than they previously could. Or people are reallocated from the AI assisted task to another. Or some combination thereof.
jjk166··on CEOs who think AI replaces their employees are just bad CEOs
> Which indicates: the management believes there are productivity gains from AI use, but adoption lags due to inertia and reluctance to change existing workflows.

Methinks adoption lags due to management's inability to align incentives such that productivity gains are rewarded.

jjk166··on Switzerland wil have a referendum to cap population at 10M
> This is never true and just economic denialism. There is a market price for labor. If there is no supply at a given price it is not evidence that a market does not exist, only that the demand is mispriced.

There can be situations where the market for a particular type of labor does not exist. Populations aren't infinite, and if there are enough good paying, desirable jobs for full employment, then there may be no one available to do a job economically.

For example let's imagine a hypothetical town where only residents of the town are allowed to work in the town, though they can provide services to those outside of the town. Let's say 100 people live in this town, and they are all doctors. There is a hospital in this town that needs 100 doctors to run. There are other jobs to be done in this town - someone needs to pick up trash, someone needs to mow lawns, someone needs to sell food, etc. Now if you pay someone a doctor's salary to pick up trash, they could potentially leave the hospital to do that job instead; but then the hospital is understaffed. Something isn't going to get done; indeed in this scenario where there are a lot more jobs to be done than people to do them, a lot of stuff isn't going to get done, no matter how good the pay is, and the jobs that are done will be insanely expensive.

In this case you would simply allow people from outside the town to work in the town, or get more people to move into town. If you scale up this scenario to cities, provinces, and ultimately nations, it's clear that at some point you must choose between structural unemployment (ie number of workers greater than number of jobs to be done), bullshit jobs (people who would be structurally unemployed are hired to do unnecessary tasks), a managed economy (employment opportunities restricted to ensure necessary work gets done at any population level), or immigration/emigration of labor (labor supply varies to meet demand) regardless of wages. In practice you'll likely get a combination of the above.

jjk166··on A Farmer Donated Land to Turn into a Park. The City Is Building a Data Center
Let's say the law is currently "this area is zoned for single family residences limited to at most two stories" and you change that law to "this area is zoned for residences limited to at most four stories." You have relaxed the law. It doesn't guarantee that you will get higher density housing, but the zoning obstacle has been removed and in many cases that is the sticking point.

Such a change does not permit a massive datacenter, or an oil refinery, or even a 5 story apartment building to be built. The laws aren't being pushed, they are being purposefully changed.

Now if some city council members get an attractive offer "for their constituents" to rezone a plot of land to build a datacenter on, that doesn't really have anything to do with the original zoning restrictions, and certainly is not affected by the YIMBY efforts. Zoning reform is not a strategy to reduce corruption.

jjk166··on India's surprise baby bust
And more specifically there is a delay between when technology lowers the infant mortality rate and it becomes culturally normal to have the smaller number of children leading to the same number of adult descendants, so you typically get 1-2 generations with way higher populations than preceding generations and then the rest after that is just this large initial population having a normal number of kids. This means you get massive populations and intense crowding decades after the fact.
jjk166··on A Farmer Donated Land to Turn into a Park. The City Is Building a Data Center
> How do you imagine YIMBY relaxing zoning and allowing things at lower cost and not simultaneously allowing people with money to push through what they want?

Relaxing zoning laws never meant throwing them completely out the window. It has always been a matter of pruning zoning rules that overly restrict land use to the point that minor deviation from the norm is impossible.

jjk166··on A Farmer Donated Land to Turn into a Park. The City Is Building a Data Center
Why shouldn't B have standing? They presumably are residents of and taxpayers to city C, and they face property devaluation stemming from nearby municipal actions.
jjk166··on I'm skeptical about efforts to revolutionize schooling
> I am 4. I have many interests. I would love to read books about those interests, but in order to do this, I have to do phonics drills and practice sounding out words. But I am 4, and I do not have the cognitive skills to force myself to do unpleasant practice to acquire a skill which I will some day cherish. I must be made to learn.

Maybe we shouldn't be forcing people to do drills and practice at a time when they lack the cognitive skills to force themselves to do drills and practice, and we most certainly shouldn't be penalizing those who struggle with such a regimen. We live in a marvelous age where you can learn about things through a wide range of media which do not require any one particular gating skill. So long as children are engaged, eventually they're going to reach a point where there are so many things they want to read that the effort to read is no longer daunting. If well structured, they'll find that in their previous learning they've actually already picked up quite a bit of understanding that helps them.

The very worst thing you can do to a child is try to shove them through a process that was not designed for them, pressure them to succeed where they were set up to fail, then tell them the failure is due to a lack of effort on their part.

The work is in setting up education programs where interest in cultivated and challenges are calibrated to the level of a student's abilities such that what they want to learn and what they need to learn are aligned. This is not easy, but life does not guarantee there is an easy way to do everything. Children are not the only ones who must learn the value of putting in the effort to reap a bountiful reward.

> That said, if left to their own devices, they simply will not do what is best for them. You have to make them do stuff sometimes, including learning.

Of course inexperienced children left to their own devices may not make the best decisions, and experienced adults must at times force them to do things for their own good. However you have to actually know what is better for them. So many terrible practices have been perpetuated because "I was ultimately better off for it." Once you accept that no one who came before you knew what they were doing, that they were all working with less information available than what you have now, and that in many cases you succeeded in spite of those shortcomings, not because of them, then you become cautious when playing the "I know better" card.

jjk166··on Iran Shock Jolts Asia and Europe to Speed Up Energy Transition
It's dropping that one last straw on the camel's back which caused the problem. This isn't an overburdened camel issue.
jjk166··on Iran stops negotiations with U.S., vows to 'completely' block Strait of Hormuz
> An infinitesimal reduction in supply of a product with a perfectly inelastic demand curve will result in an infinite increase in price.

And an infinitesimal reduction in supply of a product with a perfectly elastic demand curve will result in zero increase in price. So the real value has to lie somewhere between 0 and infinity. That's not a particularly useful prediction.

> More concretely, the oil supply tightening in response to COVID was considerably smaller than todays, yet oil prices spiked higher than we have today.

That is not what happened, oil prices tanked in response to covid despite global oil production falling by about 10%. 2020 had the lowest average price of Brent Crude since 2004, and did not recover to the level immediately before the pandemic until March 2021. You may have been thinking of the elevated prices in early 2022 shortly after the Russian invasion of Ukraine which peaked in May 2022. Of course that war which has continued for years affecting a major oil exporting nation only led to an 18% increase in price, quite a bit below the 41% increase seen in the current situation.

jjk166··on Iran stops negotiations with U.S., vows to 'completely' block Strait of Hormuz
I say it usually results in a lot less. Prove me wrong.
jjk166··on Can You Stop a Hypersonic Missile?
You shouldn't have because the poor resolution is irrelevant and OTH radar is widely used in BMD systems.
jjk166··on Iran stops negotiations with U.S., vows to 'completely' block Strait of Hormuz
By what objective standards does a price 41% higher than prewar encode an optimistic outcome?

https://xkcd.com/904/

jjk166··on Can You Stop a Hypersonic Missile?
You don't need to target at distance, you are limited by the range of your interceptor. OTH tells you which interceptor battery needs to be looking where.
jjk166··on Can You Stop a Hypersonic Missile?
Knowing that it's coming and the general direction is what you need for command and control, the actual interceptor is going to be much more limited in range than the targeting radar.
jjk166··on Can You Stop a Hypersonic Missile?
> Ground radar is line-of-sight. The Earth is round. Anything below the horizon is invisible.

That's not true. Ground based radar can see over the horizon (hence the term Over-the-horizon radar) by taking advantage of the refractive index of air, allowing the radar waves to essentially curve along, as well as by bouncing radar waves off the ionosphere.

Given that this is the foundation of TFA's argument, it does not instill confidence.

https://en.wikipedia.org/wiki/Over-the-horizon_radar

jjk166··on What appear to be biochemical processes may be a natural feature of geology
If there's enough radioactive material and it is mobile enough (due to ground water or wind driven mixing) to stay near the surface it could sterilize any organic material that comes in faster than it can accumulate.
jjk166··on A sleep-like consolidation mechanism for LLMs
> The question was "why is it needed".

No that was not the question. Why precisely sleep is essential is a complete non-sequitur to the original question, which was "does something occur during sleep which resembles what is described in TFA such that it can justifiably be called sleep."

As a general rule of thumb, if you find someone's responses incoherent, it's good practice to check what is actually being discussed.

jjk166··on A sleep-like consolidation mechanism for LLMs
It is not necessary to answer the question of what it's doing that is essential to know a subset of what it is doing.
jjk166··on U.S. Seeks to Give Weapons-Grade Plutonium to Startups for Fuel
> If it's too diluted, you can purify it with normal chemical processes and then make your bombs. It's not like uranium where you need a monstrously expensive isotope separation process to get the fissionables out.

While technically it is possible to separate plutonium out chemically, it is extremely difficult to do in practice. Plutonium separation plants are more expensive than uranium enrichment plants, and that's the reason states pursuing clandestine nuclear weapons programs choose uranium enrichment over plutonium. Uranium enrichment was much more expensive in the 40s and 50s when the US built up its plutonium production infrastructure, all of which was shut down in the 80s. Modern centrifuges make uranium enrichment cheap and simple, and the only economical source of plutonium is dismantled nuclear weapons.

jjk166··on Starship's Twelfth Flight Test
Well the idea would be to put a catch tower on a barge so those aren't actually issues.
jjk166··on Magnifica Humanitas
> We need to fix our government and then you can build anything you want.

Are we all going to stop building things until the government is fixed and all political problems are resolved? No? Then we must think about how we should build in a world with imperfect government.

jjk166··on SpaceX punts Starship launch as investigation opens into Starbase worker's death
I'm surprised mining is so low. As an outsider, it seems like similar work to construction but with the danger turned up to 11. I can think of multiple potential explanations, but I have no intuitive sense of which, if any, is likely right.
jjk166··on Starship's Twelfth Flight Test
So 200-280 tons. A standard barge can support 1500-3000 tons of cargo. Even with the added weight of a catch tower and a healthy factor of safety thrust isn't gonna sink the barge. Far more likely the major hurdle would be stability issues with how tall it is.
jjk166··on Not alive, but not dead: disembodied human brains used for drug testing
Let me add Johnny Got His Gun to the surprisingly large number of works that seem to anticipate exactly this premise.
jjk166··on Actually, democracy dies in H.R.
Please avoid strawmans when using this forum. Also the personal attacks weaken your already weak argument, FYI.
← PreviousPage 6 of 34Next →