Hacker Newsnew | past | comments | ask | show | jobs | submit | organsnyder's commentslogin

I've seen real gains in productivity using it. Nowhere near the 10x some people are promising, though, let alone replacing me.

> ...the reward of mentoring.

I really feel this. Claude is going to forget whatever correction I give it, unless I take the time and effort to codify it in the prompt.

And LLMs are going to continue to get better (though the curve feels like it's flattening), regardless of whatever I do to "mentor" my own session. There's no feeling that I'm contributing to the growth of an individual, or the state-of-the-art of the industry.


Not sure what you're implying, but I'd say their society is doing fairly well.

That's about as much as my Framework Desktop cost (thankful that I bought it before all the supply craziness we're seeing across the industry). In the relatively small amount of time I've spent tinkering with it, I've used a local LLM to do some real tasks. It's not as powerful as Claude, but given the immaturity in the local LLM space—on both the hardware and software side—I think it has real potential.

Cloud services have a head-start for quite a few reasons, but I really think we could see local LLMs coming into their own over the next 3-5 years.


I'm definitely a better driver because of bicycling. You gain new skills when you know that you're going to come out the loser in almost every collision.

If you cycle and it doesn't make you a better driver (and pedestrian) you're probably a bad driver and a bad cyclist.

I've worked with many "enterprise IT" sysadmins (in healthcare, specifically). Some are very proficient generalists, but most (in my experience) are fluent in only their specific platforms, no different than the typical AWS engineer.

Perhaps we need bootcamps for on prem stacks if we are concerned about a skills gap. This is no different imho from the trades skills shortage many developed countries face. The muscle must be flexed. Otherwise, you will be held captive by a provider "who does it all for you".

"Today, we are going to calculate the power requirements for this rack, rack the equipment, wire power and network up, and learn how to use PXE and iLO to get from zero to operational."


This might be my own ego talking (I see myself as a generalist), but IMHO what we need are people that are comfortable jumping into unfamiliar systems and learning on-the-fly, applying their existing knowledge to new domains (while recognizing the assumptions their existing knowledge is causing them to make). That seems much harder to teach, especially in a boot camp format.

As a very curious autodidact, I strongly agree, but this talent is rare and can punch it's own ticket (broadly speaking). These people innovate and build systems for others to maintain, in my experience. But, to your point, we should figure out the sorting hat for folks who want to radically own these on prem systems [1] if they are needed.

[1] https://xkcd.com/705/


I don't really think so. That was a ship that sailed ten years ago and nearly every sysadmin who is still proficient with managing on-prem stacks has adapted to also learn how to manage VPCs in an arbitrary cloud. It's not like this is a recent change.

They run fairly well for me on my 128GB Framework Desktop.

what do you run this on if I may ask? lmstudio, ollama, lama? which cli?

I run Qwen3-Coder-Next (Qwen3-Coder-Next-UD-Q4_K_XL) on the Framework ITX board (Max+ 395 - 128GB) custom build. Avg. eval at 200-300 t/s and output at 35-40 t/s running with llama.cpp using rocm. Prefer Claude Code for cli.

Curiously, how come you chose -Q4_K_XL instead of -Q8_K_XL?

Can't speak for parent, but I've had decent luck with llama.cpp on my triple Ryzen AI Pro 9700 XTs.

Is Grok driving any x.com subscriptions?

I don't have enough time for that.


My homelab is in a closet that has the water meter for my above-garage apartment. Before I put a heater in the garage itself, I needed to make sure that that closet didn't freeze. I rigged up a temperature sensor to start mprime on a server if the temperature got too low, but higher than the electric heater that's also in that closet. I figured I might as well contribute to research if I'm just burning watts for heat anyways.


Where can I buy a compute-for-heat home system?

Edit: I have a heat pump, which is more efficient for heating of course


A late model Intel MacBook from eBay and pretty much any Electron app from 2020 should do it.


I remember some startups trying to install cryptominers in people homes, the idea was to use the electricity that would be spent heating the space anyways. The company would pay for the mining hardware while the customer would provide the electricity, and the profits would be shared.

I don't know how it worked out, but the idea was there.


I know of this one [1], a 1000W space heater with integrated cryptominer. Looks kike you can actually buy it now. Not sure how much the mined crypto offsets the heating costs though.

[1] https://21energy.com/products/ofen-2


You might prefer a heat pump.


I do and have one actually. I have no idea if a kWh of compute could be worth more than eg. a kWh/(heat pump COP) though. Probably not...


My understanding is that for most residential heat pumps, the temperature needed to make the heat pump less efficient than resistive heating is so low that it enters a range that the pump doesn't even work anymore.

However, that's only a measure of efficiency. It could still be that the throughput isn't enough. A 30 kW resistive heater can ALWAYS output 30 kW of heat. But my 7 kW heat pump could produce anywhere from 14 to 30 kW depending on outside temperature.


Does that mean the heat pump gets less efficient as the outside warms? Because that would be fine. 7kW to make you home a constant temperature seems wonderful.


No, they get less efficient as the outside gets colder.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: