hackernudes

The word is terawatts unless you mean earth-based watts. OK then, it's confirmed, data centers in space!

High latency to earth but low latency (potentially) to other satellites.

DROPS, The Dresden Real-Time Operating System Project -- https://os.inf.tu-dresden.de/drops/

QNX -- http://toastytech.com/guis/qnxdemo.html (or https://news.ycombinator.com/item?id=44856046)

Pretty sure it was in Snow Crash

Neverball is a similar game that's been open source for ages. It has a web based version too: https://play.neverball.org/

The agents.md is from the upstream tinygrad repo: https://github.com/tinygrad/tinygrad/blob/master/AGENTS.md

Context: https://www.cisecurity.org/cis-benchmarks, https://www.cisecurity.org/about-us

"""The CIS Benchmarks® are prescriptive configuration recommendations for more than 25+ vendor product families. They represent the consensus-based effort of cybersecurity experts globally to help you protect your systems against threats more confidently."""

Do you still have the details details about fixing Windows 11 boot over iscsi?

I love the idea of injecting drivers into windows and did a lot of experimenting along those lines (using WinPE to install drivers offline and edit the registry).

I upgraded my home network and am booting over iscsi using ipxe. My only two remaining issues are: - Windows 11 25H2 fails to boot since 26200.7019 (and 24H2 since 26100.7019). - Windows 10/11 S3 sleep+resume does not work with WinOF-2 (mlx5.sys) driver

Everything works well in linux, though!

I am using NixOS. I had to customize the initrd to use iscsistart to connect to the target. It is also important to run iscsid when the system boots to automatically reconnect (which annoyingly takes 10-15 seconds when I resume from sleep). I am using iSER (iscsi over rdma), but TCP worked fine too. I export ZFS zvols on the server over iscsi using targetcli (which configures the in-kernel iscsi target support, sometimes called LIO).

I tried NFS but the performance was bad.

It comes down to semiconductor manufacturing, not ASIC design. Taiwan, Korea, USA are still on top.

If you have flip flops, it's not "no memory".

If you have a ROM, it's not "no memory".

Needlessly pedantic!

I thought this was pretty cool but the first video didn't play. All this write up and I really just want to see the damn demo in action first! (Edit: reloaded the page and it worked. I still would like to see it on rela hardware!)

Yes, the three playback limit is so annoying. Just... why?! CEC is so stupid. Way overengineered yet completely undercooked. I'm imagining some day soon TVs/receivers will start proxying the CEC bus instead of sharing it globally.

I wrote a program in Golang to control my a/v setup. Included within are small pkgs to control Linux CEC and LIRC devices (ioctl/read/write) as well a pkg for LG TV commands over serial port. Link here: https://github.com/EBADBEEF/tvman

One really useful thing when getting started was to use `cec-ctl -M` to monitor the CEC traffic live. Like the author, I used the v4l-utils commands to interact with CEC but eventually got frustrated with them and rewrote my program in in Go!

I have found CEC to be flaky and hard to work with. I had to turn off CEC on my TV because it breaks everything, almost randomly switching inputs and turning on and off devices.

The Scythe books are written by Neal Shusterman!

I meant that they train their system on pictures where they have the underlying medical data. Their system might still be total crap (teehee), but I'm guessing that they at least try to make it predictive/generalized.

They probably do clinical trials (or at least something like that) where they get baseline data from participants through other means.

There are two issues to consider: security updates and software compatibility.

The LTSC version is good for security updates, but I worry that software could stop supporting Windows 10 despite the LTSC version existing.

Coincidentally I am about to install Windows 10 Enterprise LTSC! I was experimenting (and struggling) with PXE boot with iSCSI. An update broke iscsi boot in Windows 11 25H2 (26200.6901 works, 26200.7019 fails) as well as LTSC (26100.6905 works, 26100.7178 fails). There were other issues with iscsi boot on the LTSC version - the network hardware needs to be enumerated before the first boot, but can't boot because it needs network (a chicken-and-egg style problem).

I 100% agree with this sentiment

I wonder if the compiler could still be a native binary (but still producing code for the target architecture). The motivation is to have the performance of a cross compile with the simplicity of a native build. I had that idea a long time ago but never tried it.

CZ already served the time in prison. It's not clear to me whether he and Binance have paid the fines yet.

Agree. Makes sense for a mouse cursor but not touch.

I've had similar problems but no amount of tweaking vm and vfs cache settings helped. Swap or not, both 32gb and 128gb of ram. Manually reclaiming memory would un-lock the system (/sys/fs/cgroup/memory.reclaim).

I wrote a user space memory reclaimer and have not got a lockup since. https://gist.github.com/EBADBEEF/f168458028f684a91148f4d3e79...

The Lightning Network (an open payment layer built on top of Bitcoin) or some other cryptocurrency.

I don't think it is macabre at all. "Hang" just means suspended. A ball on a string is hanging instead of falling. Laundry can be hung out to dry. It's about something that would normally move that is not doing so. Also see "hung jury".

My understanding is this is for excess PV capacity. But I think they should address this! Their plan is to heat deep inside a big pile of dirt using resistive heating elements and then pump the heat out in pipes. Maybe it's hard or expensive to add solar heating into the system? But why couldn't they pump it in? Maybe they don't have extra space because the dirt mound is covered in PV panels.

The article is about accessing a service (nostr) through a hosted web app. The domain or server that is hosting the app could be compromised and serve a bad app.

Posts on nostr use a key pair so when you see a post from foo you know it's the same foo you knew from last week. Also, posts are shared to and stored on multiple independent servers (called relays).

A compromised app could serve you fake posts or censor stuff.

Google is famously a monorepo and is basically the gold standard of CI/CD.

What does happen is APIs are constantly upgraded and rewritten and deprecated. Eventually projects using the deprecated APIs need to be upgraded or dropped. I don't really understand why developers LOVE to deprecate shit that has users but it's a fact of life.

Second hand info about Google only so take it with a grain of salt.

Looks like a flutter app that can build for desktop Linux https://github.com/parres-hq/whitenoise_flutter .

I started my reply thinking it was still using Tauru but apparently things change fast!

Are you using cloudflare DNS? https://news.ycombinator.com/item?id=19828317

Why do you say that? They are on their seventh iteration of hardware and even from the beginning (according to the article) they were designed to serve Google AI needs.

My take is "sell access to TPUs on Google cloud" is the nice side effect.