brunohaid

Was concerned about splinters too, but if you compress and treat/seal it a bit, it works quite well, although they do get a bit gross after 2-3 years to the point where you should flip/replace them.

It took my about 3 days to get over it after seeing it on the final images.

If you swear at them long enough they start disappearing...

Mostly bottom of back shelves and behind vertical shelf beams.

Thanks! Had a really good camera & lenses for the day (Hasselblad X1D II with Zeiss Otus and Canon TS-E 17mm), and then post was mostly reducing colors besides brown/green/red hues a bit.

Thanks! And: Same. It got much better since the early days of MSFS 2020, and tried picking gear that has solid drivers/good scripting APIs, but I also rarely get it out just for fun, for the same reasons (except the occasional barrel rolls in a TBM over Manhattan cravings). It is great to have for a bunch of practice runs before actually heading to a new airport for the first time though.

None taken! Cut the shelving to my (elbows at 90 degrees) standing desk height and used it that way about 80% of the time, and for the rest the drafting height barstool like chair worked well, as the legs are naturally angled.

Took a bit of shuffling around, biggest challenge was getting a somewhat decent frame given the incredibly small space - main camera facing the window had simple Sigma MFT lens that worked in the end, with a hood on to make sure the ring light doesn't diffuse into it.

Is it the line length that's tedious, or the columnar layout itself confusing?

Most common term in the US is something like Boltless Shelving, specific shelves used here were https://www.globalindustrial.com/c/storage/shelving/boltless...

Oh, didn't think that'd make it to the front page, appreciated! OP and builder here.

The website was purely because a friend and I were looking for design work during lockdown and put together a couple of things we recently worked on, but basic design and build was a fun ~6 months solo project.

We had a good discussion on https://www.reddit.com/r/architecture/comments/1mlo6hu/tryin... over the weekend with more details, but also happy to answer any questions here.

For me it's a combination 1) solid foundational choices all along, no bolted on vanity features or constant rewrites chasing the latest trend, with everything well documented and 2) incredibly responsive founding team, so you get very quick answers from the people actually building it.

Good point and was mostly re Typesense (can't edit the comment anymore).

But given that duckdb handles "take this n GB parquet file/shard from a random location, load it into memory and be ready in < 1 sec" very well I'd argue it's quite easy to build something that scales horizontally.

We use it for both the importer pipeline that processes the 2B row / 200GB compressed GBIF.org parquet dataset and queries like https://www.meso.cloud/plants/pinophyta/cupressales/pinopsid... and the sheer amount of functions[1] beyond simple stuff like "how close is a/b to x/y" or is "n within area x" is just a joy to work with.

[1] https://duckdb.org/docs/stable/core_extensions/spatial/funct...

Bit thin on details and not looking like they’ll open source it, but if someone clicked the post because they’re looking for their “replace ES” thing:

Both https://typesense.org/ and https://duckdb.org/ (with their spatial plugin) are excellent geo performance wise, the latter now seems really production ready, especially when the data doesn’t change that often. Both fully open source including clustered/sharded setups.

No affiliation at all, just really happy camper.

Hand curated by topic, and then simple admin overlay to hide posts that are promos etc, so a bit of curation there.

There are some feeds that sometimes have great content, but also bunch of promo etc noise, so it's a bit tricky to decide who makes the cut.

Even in the year 2025 of the great enshittocene, if you put something on the internet, it should be possible for people to consume it the way they want. Including the very simple question: what‘s new since I last checked?

This example is for plant/fungi/biodiversity nerds, but maybe also a template for others to build more bridges in and out of algorithmic monopoly silos. To, y‘know, make the world more open and connected.

Definitely doable, aquarium controllers are also quite flexible.

We're working on a free Home Assistant plugin that let's you define your sensors and outputs, then choose a plant (see link below, or climate of a specific location) and hopefully take most of the heavy lifting from there.

LEDs help to supplement in regions that don't have that much light (Netherlands etc) or when you try to achieve something very specific (scientific, flavor profile etc) but bypassing the reactor in the sky is just not a very good idea.

The other issue you mentioned also killed 50%+ of vertical farming startups: you need a diverse living ecosystem. You need a living soil crawling with organisms of all sizes (check Arbico Organics, Jim's Worms and others), the whole thing needs to be balanced, otherwise it's hard to grow in the first place and/or a single unpredicted "intruder" will wipe out your entire operation (Bowery Farms and plenty of others from what I understand).

Ha, working on exactly that at eg https://www.meso.cloud/plants/magnoliophyta/malvales/magnoli...

True on both counts, was a quite early comment and initially thought they're coming back in to land vs barely having taken off. Only leaves control inputs but given how short the video is it could also be that there simply wasn't much to input/correct anymore despite the slight rocking.

Can't edit anymore, but the general gist of catastrophic failure needed to prevent a 787 from climbing out of this situation still holds.

We took off ahead of them.

Good question. False sense of routine, experience? Definitely a pet peeve.

Was flying as a passenger on a really small airline (8 seater plane, Green Air) out of San Jose in Costa Rica. We got cleared for takeoff ahead of a United 737, at most 500 feet into the humongous runway for that plane. Yet the pilots still put in the 2 minute effort to taxi back to the beginning of the runway, even though they could have easily taken off from where we entered it. Don’t know if it was their protocol or the pilots decision, but I will trust this airline for a very long time.

„Runway behind you“ is drilled into your head as one of the useless things in aviation. You always want to make sure to use the most runway available to you, exactly for cases when something happens. Hypothetical in this case: you realize something’s wrong with the plane, but you’re already too fast and close to the end of the runway to reject the takeoff because you wouldn’t be able to stop in time anymore.

Large airports with heavy traffic sometimes have operational constraints to send a plane out ahead of another from some intersection, but if the ADSB data is correct, taking off from half the available runway in a fully loaded 787 isn’t a good idea. You just give up a ton of margin for errors.

Baffling tragedy, again.

Clear skies, no LiveATC but reports of single Mayday call, gear out but no flaps and no control inputs visible in the grainy video. Something has to go really catastrophically wrong with a modern jetliner for that to happen, like the very dense flock of birds in Korea with the 737 a couple of months back.

The very short intersection takeoff seems like a good hint (and terrible practice), but all gears and engines look kinda OK from the outside. If they‘d scraped something on takeoff hard enough to take out both engines, there’d probably be some visible damage, or at least some gears sheared off.

EDIT:

Fully agree with the speculation in light of tragedy comments, but aviation is a bit of a special case. The reason it’s so safe is because an awful lot of people immediately start looking into potential reasons and then spend years getting to the bottom of it. The initial speculation is like an exercise: what could have happened? What if I’m in that situation, and need to act now, without knowing much of anything? If you do that a couple of dozen or hundred times throughout your life, it really builds a foundation for when an actual emergency ever happens to you.

It’s a bit like the reason most flight attendants in the emergency exit jump seat across from you won’t talk with you during the actual takeoff and landing: they‘re mentally walking through a potential emergency and what they‘d then need to do. Every single time. So if it ever happens, there‘s muscle memory, 10000x over.

EDIT 2: see the Flightradar24 comment below, it looks like they did backtrack and use the full runway.

You might want to look at https://typesense.org/ for that.

Thanks for the excellent comment & insight!

Appreciated! Didn’t know Cline already does RAG handling, thought I’d have to wire that up beforehand.

Noice!

Does anyone have a good recommendation for a local dev setup that does something similar with available tools? Ie incorporates a bunch of PDFs (~10,000 pages of datasheets) and other docs, as well as a curl style importer?

Trying to wean myself off the next tech molochs, ideally with local functionality similar to OpenAIs Search + Reason, and gave up on Langchain during my first attempt 6 months ago.

Call me naive, but I somehow trust them to deliver in time/specs?

It’s also a more general comment around „AI desktop appliance“ vs homebuilts. I’d rather give NVIDIA/AMD $3k for a well adjusted local box than tinkering too much or feeding the next tech moloch, and have a hunch I’m not the only one feeling that way. Once it’s possible of course.

Still surprised that the $3000 NVIDIA Digits doesn’t come up more often in that and also the gung-ho market cap discussion.

I was an AI sceptic until 6 months ago, but that’s probably going to be my dev setup from spring onwards - running DeepSeek on it locally, with a nice RAG to pull in local documentation and datasheets, plus a curl plugin.

https://www.nvidia.com/en-us/project-digits/