loehnsberg

6 ms·

31.
by loehnsberg 1y ago
Did they use the online Deepseek Chat or the open source model. If you ask either about the Tianenmen Square you get very different answers, which may be true for response quality as well.
32.
by loehnsberg 1y ago
Interesting to learn that o4-mini-high has the highest intelligence/$ score here at par with o3-pro which is twice as expensive and slow.
33.
by loehnsberg 1y ago
I made that same experience. You can get individual people to install it, sure, you may even get a group to do this if you start it, but good luck convincing the other parents of your kid‘s sports team group, which btw you must be part of.
34.
by loehnsberg 1y ago
L1-regularized autoregressive features, holiday dummies, Fourier terms (if suitable in combination) yield lower test errors, are faster in training, and easier to cross-validate than Prophet.
35.
by loehnsberg 1y ago
Check out the Mittelmann benchmarks: https://plato.asu.edu/bench.html
36.
by loehnsberg 1y ago
I would argue that not spending money on it and showing to upper mgmt that the folks they hired can actually get the job done often contributes to an external contractor not getting hired. You might want to question, why didn‘t you ask the
37.
by loehnsberg 1y ago
Assuming you add all the annoying details that algo trade execution brings, the algorithm still provides the answer on which position to take within a few microseconds, which is what you want if you trade in a limit order book.
38.
by loehnsberg 1y ago
There‘s still a lot of work left to do to go from an academic prototype to live trading: real-time data, market access, SCADA, compliance & legal, security, … Also you must be a physical player that owns the battery and/or right to
39.
by loehnsberg 1y ago
You get electricity at the same price in Japan and at half the price in the US. Message to the policy-maker: just be resource-rich. How exactly do you implement this?
40.
by loehnsberg 1y ago
Uniform pricing creates huge windfall profits for those cheap generators too, which incentivizes more investment. I‘m not saying there‘s collusion but the incentive to keep a gas turbine in the stack is clearly given.
41.
by loehnsberg 1y ago
The La Marzocco espresso maker at $7000 is cheaper than buying a new one.
42.
by loehnsberg 1y ago
If Valve was serious about Mac as a platform, they would port their Source games, like they ported Portal 1&2 for the Switch. To some extend I understand this, because Apple isn‘t serious either, no matter how often they show me a Dual
43.
by loehnsberg 1y ago
I did not customize bangs and the standard lenses are already useful. I have two custom lenses for travel and academic papers, each searching specific domains. What I meant is that you learn how to use them. What is there by default is alre
44.
by loehnsberg 1y ago
I use both but cancelled my Perplexity subscription. Kagi is the better version of Google search, especially if you learn how to use lenses, bangs, and all these features. Kagi Assistant is great if you‘re happy with basic no-frills chat, i
45.
by loehnsberg 2y ago
Why skip Xen? Xen in Black Mesa is really good. I found it to be much better than the original.
46.
by loehnsberg 2y ago
You can plant trees (or any plant really as long as they grow fast) and then bury it so that the carbon won't get released or at least very slowly. There's an older thread discussing this idea [1]. CCS would dispose the CO2 deep u
47.
by loehnsberg 2y ago
I don‘t understand why this is bad. Aggregation is a feature because it allows to be no worse than Google, in my experience likely better, and you pay for search with money and not with data.
48.
by loehnsberg 2y ago
For all those screaming confounds, you can run experiments. Just please keep sampling until you get good results for milk chocolate as well.
49.
by loehnsberg 2y ago
As they are using blocks (that cover a square area), I don‘t understand why use MC when you can actually compute the ratios directly based on the area covered by blocks that are less or equal some distance from a defined center - other than
50.
by loehnsberg 2y ago
What‘s most striking is that a sober dialogue on opposing views/ideas has been replaced by partisanship and hatred of the othet side, whatever the subject. What do we need to do to get out of this mess?
51.
by loehnsberg 2y ago
Largest applications may well be in power systems (economic dispatch, unit commitment), material requirements planning, transportation networks, but linear programming can also be used to fit functions, think constrained regression with L1
52.
by loehnsberg 2y ago
[At this very moment, we all have microplastics coursing through our bodies. This is not the fault of not enough recycling. This is the fault of too much plastic. So I say: Let’s treat plastic like the toxic waste it is and send it where it
53.
by loehnsberg 3y ago
I‘d love to see an IMDB widget.
54.
by loehnsberg 3y ago
I‘ve been using it for two years as my primary search engine, and they’ve been nailing it. Kagi‘s worth every cent. No ads, no need for DDG g-bang, because Kagi search results are above par, excellent support via email/feedback, „quick
55.
by loehnsberg 3y ago
Depends on what platform you use - iOS is quite buggy whereas I have close to zero issues on Mac. Turn on compatability mode (left button next to address bar) if a site does not work properly which turns of ad blocker etc.
56.
Show HN: Manganite – Quickly turn Jupyter notebooks into web apps(github.com)
4 points by loehnsberg 3y ago|0 comments
57.
by loehnsberg 3y ago
I think this is a great idea! My kids use iPads and managing their screen time works well with what iOS has on board, but I cannot control social media, as this would nearly always intrude the privacy of the kid.
58.
by loehnsberg 3y ago
We homeschool our two kids for two years now, because both had made bad experiences in the conventional school system. It took us a while to find our pace, but it is going very well so far. There is plenty of homeschooling material out ther
59.
by loehnsberg 3y ago
Most triangulation algorithms are designed for the two-dimensional plane, so this only works if the first two components are sufficient.
60.
by loehnsberg 3y ago
Unfortunately, as you move into higher dimensions, these algorithms typically bog down. Suppose that you have a point that is inside of the convex hull of the mesh that you want to use for triangulation (we‘re talking hyper-triangles here).

More ›