kphorn
kphorn on Hacker News
The good news is bigger laptops coming tomorrow in the phased release.
We are in the days of building the things we dreamed about as kids. Looks great for right turns!
If the flying version is ever sold (unlikely) it will be subject to flying thing rules (FAA). Best case it would be like buying a used F1 car from Ferrari. They retain possession and you get to occasionally fly in on the private jet and drive it under supervision on a racetrack.
It’s cool tech demo. I’m not expecting to see one fly in private hands ever.
I'll introduce the concept of Hobbies for those who crave the details. Hours spent in the basement/attic/office/garage/workdesk sweating the details in an inefficient and un-economic way. Learning music, painting figurines, woodworking, playing with microelectronics, making sculptures, model aircraft, ham radio. Offline things, intentionally far away from the frontier that bring joy, depth of understanding, and are intentionally not going to be gobbled up by an LLM.
One suggestion for the feeling that has been great for me. Finding a hobby where you can focus on sweating the details. Detach it from economic imperative (the core definition of a hobby, you dont make money from it). Ham radio, maker and micro electronics, woodworking, calligraphy, music. Of course there are more technically advanced solutions to the problem, that's not the point. The point is the effort and discovery.
Great blog post by Peter Gibbons, Michael Bolton, and Samir "Nagana-work-here-anymore" about their time at Initech
https://www.youtube.com/watch?v=yZjCQ3T5yXo
"all right so when the subroutine compounds the interest, it uses all of these extra decimal places that just get rounded off. So we simplified the whole thing, we round them all down and drop the remainder... in to an account that we opened"
This is an awesome technical demonstration. I enjoyed learning by doing and I’m impressed by how well this ran on a phone
This is the biggest anticipated benefit of SMR. The US has 30+ licensed reactor designs. France has 3. Korea has 3. The economics of French and Korean reactors, built repeatably, are drastically improved over US reactors. If the NRC does its job and actually says "we know more about nuclear power than the local state energy commission and operator" then we the US can achieve those levels of repeatability and cost. We want safe reactors, we dont want infinitely customized reactors that are tailored to every state and operator's preference for how they want to polish fittings and lay out the pipes etc etc.
Agree that if you run them for 40 or 60 years (1 renewal + 1 extension) it gets very affordable. Similar to data center, the economics are great if you take out construction costs. The problem is 40-60 years has proven to be a very long time in political and economic contexts, so on a practical basis many many plants are shut down prematurely and dont fulfill that useful life, increasing the cost of capital (bonds etc) to build the new plants and therefore LCOE.
Totally agree that if you get a stable, reliable operating reactor it's very cheap. When people, politics, and the rest get in the way the actual costs drastically increase.
I have the laptop version 9955HX3D (which I believe is a 9950 with 2.5ghz base, 5.5 boost, and lower TDP) in a Legion Pro 7 and it's a fantastic chip for a computer that I can move around. 16 full cores in a laptop is a luxury and 64gb of ram has proven to be a wise choice prior to price spikes
Let's not discuss battery life and overall size, however it helps the thermals.
I do a lot of CPU bound engineering and scientific compute on the computer and it chews through it.
I also bought the entire computer for $2500+tax new from Microcenter in December and it's appreciated 25% in value since then.
I disagree - Fable melts the GPUs and they have a high incentive to move people off of that. If they have meaningfully decreased cost to serve on Opus 5.5, they can reduce prices and increase margin or at least turn off the most expensive compute.
Good data and goes to show that Fable is melting the GPUs and is priced accordingly. I'd guess that cost to serve for Opus 5.5 is meaningfully lower through architecture advances
They store the empty tubes, why not the full ones? Are they collecting that much mass that it’s really impractical. Are the empty tubes nested like solo cups to save space? I understand the rationale for the sample return, but those tubes will be hard to find after years of dust storms.
Agree, the decision to drop samples on the ground doesn’t make sense to me. Store them in the rover. Much easier to find them all in one spot later.
I lived and worked in a van full time for a while. There are easier ways to save power than delidding a MacBook. Like using a 12v USB-PD adapter that connects directly to the battery.
I suppose there’s a particular attachment to the “Low power display” in this article but that kind of runs against the whole van life argument which is live light and have a small footprint.
Good news that the new model is the "Most capable, most aligned model".
The risk hasn't been stated clearly - it's now a classic arms race.
A well-resourced organization trains their own, highly persistent, highly-capable, safeguard-free, and unaligned model and deploys it on 1000x GPUs with a message board and a nearly-impossible objective. No infrastructure is safe. No organization is safe.
You need your own 1000 bot swarm to scan, identify, and defend against the threat, which means investing in infrastructure and capabilities to defend. Cost and complexity go up. Risk and attack surface goes up.
The price reduction has absolutely tanked the resale market for these. Thanks a lot Tim Apple
It’s more AI copypasta as it goes further along where it gets really into the how and not the why.
“But here’s the twist: that search I described earlier was never sent to botd. All the history was in SQLite, so I pointed another agent at it and got the analysis anyway. The tool died; the data didn’t.”
If I had more time, I would have written a shorter letter…
It’s basically impossible to compete on economic terms with deeply subsidized hardware that is widely available to rent or as a service with zero commitment.
For general inference there’s no ROI that makes this work vs subscriptions.
25k for computer now, plus 9-10% sales tax, plus operating cost, plus time and cost for R&D tinkering with models, harnesses, and infra (assuming highly capable engineering talent that can get paid for your human inference) vs a HEAVILY subsidized subscription at 200 per month with free R&D has a pretty long ROI (15 years?)
At API costs, it’s like 6 months if you’re heavy on inference.
For training, specialized models will have their own ROI that makes this worthwhile. Then debate renting capacity and the platform to choose
Yes, the term of art is “resources”. Human Resources
We are onboarding resources
We are cutting resources
We are shifting resources
The Civil Reserve Pizza fleet always amazed me. About once a quarter, the Air Force Academy would get dominoes pizza for all 4000 cadets. every cadet + staff and some visitors gets a medium pizza for dinner. that's about 4000 pizzas stuffed into the back seats of honda civics from around colorado springs arriving in a 2 hour span. I dont know how they did it but boy could they mobilize
I think the only thing that the human did was remove the emdash between the two sentence fragments and replace with a period.