fnands

5 ms·

31.
by fnands 6mo ago
They do some good tours. They are first come/first serve, so show up in the morning for the best chance to get a place.
32.
by fnands 6mo ago
Nice, I didn't get it. I just realized "Kyouryuu" is Japanese for dinosaur. And Pfirsich is of course Peach.
33.
by fnands 6mo ago
They got the intern to write a funny post. Probably took them an hour and it drums up a bit of good press for the Hi-Lumi LHC. Money well spent if you ask me.
34.
by fnands 6mo ago
I always look forward to CERN's April fools jokes!
35.
CERN levels up with new superconducting karts(home.cern)
395 points by fnands 6mo ago|85 comments
36.
by fnands 6mo ago
Could be as if you live in a city in Germany you don't necessarily need a car? But I would be surprised if car seats are the main driver here in any case.
37.
by fnands 7mo ago
Related (OpenAI announcement): https://news.ycombinator.com/item?id=47438716
38.
by fnands 7mo ago
Woah, first Anthropic buys Bun, now OpenAI Astral? Seems like the big AI players love buying up the good dev tooling companies. I hope this means the Astral folks can keep doing what they are doing, because I absolutely love uv (ruff is pre
39.
by fnands 7mo ago
> counted my house as a tree Even the more sophisticated algorithms pretty much always do this ;-) You are probably not interested in taking this further, but you could give the Li tree filter a try: https://pdal.org/en&#
40.
by fnands 7mo ago
I gave a talk about the paper in our internal journal club recently (we work on similar problems, usually using stereo imagery though). It's a nice piece of work. I especially like the sections on data cleaning and registration, as tha
41.
by fnands 7mo ago
You could look it up in the metadata file: > We additionally release a global GeoTIFF of input image acquisition date, where pixel values encode year minus 2000 (e.g., 18.25 indicates April 2018) That being said, I am sceptical on how ac
42.
by fnands 7mo ago
It's from FAIR, i.e. their fundamental research arm. Maybe there are some ulterior motives, but they do also just do a little bit of "feel-good" research. This was also in collaboration with the World Resources Institute and
43.
by fnands 7mo ago
The USGS Lidar data is a treasure trove, I use it a lot at work. What did you do to actually count trees? Even from aerial Lidar it can be a bit finicky for closed canopies.
44.
by fnands 7mo ago
From the paper: > CHMv2 is derived from single-date imagery, where the acquisition process selects the best available image within a target period (2017 -2020). This limits the direct use of the released CHMv2 data for attributing canopy
45.
by fnands 7mo ago
> 12 hour ahead prediction As the comments on that article rightfully point out, restricting the data analysed to 12 hours before the resolution feels like cherry picking.
46.
by fnands 7mo ago
I think there are lots of plots that revolve around rigging sports betting (e.g. sub-plot in Pulp Fiction), but I can't really think of a case where it is not so much about manipulating the actual act that occurred, but rather the news
47.
by fnands 7mo ago
At least cryptocurrencies had some nice ideas behind them. Just sad they almost immediately got co-opted by swindlers and criminals.
48.
by fnands 7mo ago
Ugh, the sophistry (or at least self-deceiving arguments) people throw around to defend these markets makes my stomach churn.
49.
by fnands 7mo ago
Man, something like this is going to be a plot point in a movie/tv series soon. Could work in some crime procedural.
50.
by fnands 7mo ago
This is the way
51.
by fnands 7mo ago
Probably just a satellite office. Might be to be close to some of Yann's collaborators like Xavier Bresson at NUS
52.
by fnands 7mo ago
Some of those pictures are delightfully cursed
53.
by fnands 7mo ago
Yeah, either that or the commoditization of LLMs. I.e. in the early days (2023-2025, lol), companies were mostly differentiated by the quality of their LLM, i.e. GPT-3.5 was leagues ahead of anyone else, so was for serious work the only gam
54.
by fnands 7mo ago
I agree that "full engineering" was a bit broad. I should probably have said something like "agent-only coding"? I.e. the point where the agent writes all the code and you just verify.
55.
by fnands 7mo ago
I'm an ML engineer, so it's mostly been setting up data processing/training code in PyTorch, if that helps.
56.
by fnands 7mo ago
When was the last time you tried? I think trying agents to do larger tasks was always very hit or miss, up to about the end of last year. In the past couple of months I have found them to have gotten a lot better (and I'm not the only
57.
by fnands 7mo ago
You'll probably be one of the best paid people in Thuringia ;-)
58.
by fnands 7mo ago
LiveEO | Senior ML Engineer | Berlin, Germany | Hybrid | Full-time LiveEO leverages high-resolution satellite imagery and AI to provide actionable insights across industries—like protecting power grids, monitoring critical infrastructure, a
59.
by fnands 7mo ago
The link seems to have the wrong text: * Proven leadership in security for modern cloud-native systems and products. Did you copy paste a security position?
60.
by fnands 7mo ago
Oh hey, the man himself! I was looking at the zlib-ng crc32 implementation which is where I saw that it was recently updated to include your algorithm. Good work, it's a surprisingly elegant solution when compared to the braiding appro

More ›