microtonal
Joined 16 years ago
Karma
26,567
Submissions
6,774
Member since
Jun 29
Opinions are my own.
View on Hacker NewsRecent submissions
Comment · 2 hours ago
You would be surprised. A lot of cheap consumer routers are have two or four ARM cores. The bigger issue is that they are typically running a kernel from the stone age (< 6.0.0). Sometimes a still-supported LTS kernel like 5.10 or 5.15, but then they stopped at an old patch level.
Heck, apparently even the new Unifi UCG Fiber is running 5.4.213 (September 2022), but maybe they backport some fixes?
[1] https://community.ui.com/releases/dd0283dc-bf71-4140-ae1d-c1...
Comment · 9 hours ago
That doesn't address the criticism that this thread started with:
Unified Push, which works flawlessly on Molly, could easily be implemented in the regular client.
Comment · 1 day ago
Thanks. I misused the idiom as a non-native speaker, well try to remember next time!
Comment · 2 days ago
Not to downplay his accomplishments (I have sung my praises in another comment), but even vector models were already around and somewhat popular before. Much of it started in Information Retrieval with things like TF-IDF vectors and later PPMI matrices that were converted to small, dense matrices with SVD. Of course, later Yoshua Bengio et al. introduced neural language models for training embeddings.
There is even a paper by Levy & Goldberg from 2014 that shows that skip-gram with negative sampling implicitly factorizes a PPMI word-context matrix.
But Mikolov's formulation is very nice in that it is a simple linear model that allows for a very simple Hogwild optimization that worked very well/fast on multi-core CPUs of the time.
Comment · 2 days ago
Well-deserved. Even though more complex language models won in the end, his stripping down of word embedding training from RNNs/convolution models to a simple linear model was the thing that was needed for compute at the time and completely dominated the field for five years or so. I do miss the times a little where you could train a cutting-edge model with a reasonable CPU (for training embeddings with word2vec) and a single GPU (some BiLSTM or transformers before BERT).
If I remember correctly, I read somewhere he initially had a hard time getting the word2vec paper accepted at a conference.
Comment · 2 days ago
And the first iteration did not have a physical Esc key (which they brought back later).
Comment · 3 days ago
Maybe you have another curl in your path?
$ uname -a
Darwin MacBook 27.2.0 Darwin Kernel Version 27.2.0: Wed Oct 14 23:47:21 PDT 2026; root:xnu-13432.41.4~25/RELEASE_ARM64_T6030 arm64 arm Darwin
$ which curl
/usr/bin/curl
$ curl https://dv-g3b-emudhra.we-have-a-problem.com/
Your client successfully connected to the site. That's bad!Comment · 5 days ago
I think it’s not worth it to them as long as 99.9% of the users hook up the TV to their home WiFi. Connecting to other networks without user confirmation probably opens a can of worms in many countries.
Comment · 6 days ago
Our kid has certainly not had access to tablets/phones the first years of her life and only rarely until she was 9 or 10 or so. We have also banned phones from the dinner table, etc. from the beginning.
Parenting takes time and focus. Putting 0-6 years olds behind a tablet screen is just absolving yourself from your responsibility as a parent. Some TV time can be ok after 3 years or so, but you have to time it strictly (so that it does not become a convenience) and exclude some types of cartoons [1].
[1] https://thesciencesurvey.com/editorial/2024/07/07/the-proble...
Comment · 6 days ago
Disconnect it from the network and hook up a streaming device that is relatively trustable (e.g. Apple TV or an open source alternative).
Comment · 6 days ago
I wonder if LEGO is going to follow the same fate, to no longer be of interest to Roblox and Minecraft obsessed kids, to develop an increasingly elderly customer base, ending up with just retirees.
I don't think this is true at all. Our daughter played a lot with Duplo and then Lego from ~2-12. Overall, she must have collected thousands of Euros of Lego from birthday, Christmas, and Sinterklaas presents. When friends were over, they usually wanted to play with Lego.
Us older people who remember the days of Lego without franchises and a focus on your own creativity/imagination might be grumpy about modern Lego (though they have been good at making adult Lego sets as well). But Lego is still wildly successful because they found ways to connect with younger generations and the franchise-based Minecraft/DC Comics/Star Wars/Harry Potter/etc. Lego is exactly one of the ways they have managed to do this, along with Lego for specific focus groups (e.g. Lego Friends).
Comment · 6 days ago
The daily fine is so meaninglessly small ($99,000 per day) that there is very little incentive to do anything.
Comment · 6 days ago
I think it's more that they happen to be aligned. Left-wing parties are more likely to tax and regulate Meta than radical right parties. So it's for their own profit. But Trump is also mostly there for enriching the oligarchy, he often has the same priorities.
Comment · 6 days ago
They also repeatedly suspended Dutch LGBTQIA+ accounts:
https://nltimes.nl/2026/04/23/meta-suspends-dozens-dutch-lgb...
It's time countries started suspending Facebook/Instagram for meddling in national affairs.
Comment · 7 days ago
Galaxy Watch Classic models still have a wheel and it's a very nice way to flick through interface items, especially in situations where a touch screen is not really accurate (e.g. when cycling or running).