ece

5 ms·

1.
by ece 2mo ago
The only benchmark that matters anymore is minimizing felonybench. Ok, long horizon benchmarks still matter too.
2.
by ece 2mo ago
Fairphone is on their 6th generation now, and in the US on Amazon. Easy repairability is as much an engineering feat as adding a button or light bar. Their support window is as long as anyone else. I'll even mention Essential who did d
3.
by ece 2mo ago
Samsung and Xiaomi have bigger ecosystems than Google in some ways. If they want to fix their development and support for issues like this, they can.
4.
by ece 2mo ago
This would be a bit convincing if there weren't other hardware and services providers like Fairphone, GrapheneOS and Google themselves, who do support and maintenance for longer.
5.
by ece 2mo ago
You'd almost think supporting a phone for a longer amount of time might actually be better than trying to sell a new phone every year or two.
6.
by ece 2mo ago
If bigger, better models aren't also leading to better safety, something is broken. Models seem to no longer argue with people, or output verbatim text, maybe the next step is cheating less.
7.
3-6-3 Rule(en.wikipedia.org)
3 points by ece 2mo ago|0 comments
8.
by ece 3mo ago
I've learned to be wary of these cars in the bay area, they seem to cut it close often, and they get stuck because they're not L5. I hope municipalities know what they're getting.
9.
by ece 3mo ago
So what's with the increased variance since 2021? Better tracking of pages maybe...
10.
by ece 3mo ago
Those scrapers can be blocked. Some scrapers won't ignore the responses, and maybe it'll lead to a meaningful reduction in scraping traffic.
11.
by ece 3mo ago
Every website that wants to, can just charge for dumps. Anubis or even Cloudflare can handle the rest I think.
12.
by ece 3mo ago
This would be interesting on a Spark or two. I don't know how seamless the GPU and networking cluster setup would be when resource sharing between agent and post-training VMs.
13.
by ece 4mo ago
What these people want is same thing people in every other field want too, which is being able to use AI as a tool and nothing more. Off course, some of the people that signed this might not recognize this, and the people who didn't si
14.
by ece 4mo ago
"It is difficult to get a man to understand something..."
15.
by ece 4mo ago
Yeah, this does seem like a needless omission. Cardboard cases work sometimes. Plastic cases aren't too hard to find, but they do require modifications.
16.
by ece 4mo ago
Controlling how to train, not over fitting, making sure training data is good, and being able to do distributed training are all reasons why open weight isn't like open source. You can't fork during pre-training with open weight-o
17.
by ece 4mo ago
Pinecube from pine64. https://pine64.org/devices/pinecube/
18.
by ece 4mo ago
https://archive.md/voqjr
19.
by ece 4mo ago
Start your own openrouter.
20.
by ece 4mo ago
Just switched to AI Plus from Pro, seems like I won't be missing much.
21.
by ece 4mo ago
72% revenue decline from a 54% price increase, but how did profitability change? Some of the features of the Deck are still differentiating like the OLED screen and touch pads, and maybe some of the price increase is meant to keep it profit
22.
by ece 4mo ago
I disagree with the first half quite a bit, COGS ultimately depends on the use case. If someone just wants something that a smaller model can do, running a local model on phone is going to have a negligible cost close to running any other p
23.
by ece 4mo ago
This is the new version, the old one is called sultan laxeby. Easily 1.5x the amount of time. PDF for assembly is still up on the website: https://www.ikea.com/nl/nl/assembly_instructions/sultan-laxe...
24.
by ece 4mo ago
Is there a bsky or mastodon extension for serving git repos and commits? Would be interesting to see this.
25.
by ece 5mo ago
Machine learning layer cake with some brute force crumbs.
26.
by ece 5mo ago
It's hard to sympathize with Anthropic for this or the export ban, the hype over model capabilities probably fuels both things (in some ways). Training data for me, but not for thee (at any scale) doesn't seem like a tenable posit
27.
by ece 5mo ago
Out of all the possible permutations of display/sensors/processing for AR, I would like to see eyewear with just a Bluetooth-like display and camera/mic. Let me use the phone or watch for all processing. Bonus points for eye
28.
by ece 5mo ago
If large models were primarily used for distillation to make smaller models better, that could be a more efficient use of resources. Smaller providers could choose to fund a larger model and benefit from it.
29.
by ece 5mo ago
Compared to an ICE, maintenance of brushes is still quite an improvement.
30.
by ece 5mo ago
Neural scaling laws are alive and well for open models, not so much for closed models when it comes to uses the general public might care about.

More ›