HN user
The best bit of that is:
In this role, you will own end-to-end bill run execution across all AWS partitions, drive the technical vision for autonomous billing operations, and build the team that ensures every customer receives an accurate cost estimated in minutes ...
Most likely Hydrazine propellant tanks:
https://www.space-propulsion.com/spacecraft-propulsion/hydra...
The current recovery status:
The full comment:
“Biological limits are real, but digital potential is infinite. If we starve our data infrastructure of cooling resources just to sustain baseline human comfort, we are actively delaying the birth of a super-intelligence that could solve all of our resource problems in the first place. Sometimes you have to prioritize the intelligence that will save us over the biology that slows us down.”
As someone spending time daily irrigating food crops I think it's time he was buried head-first in the Sahara desert for a week!
Context: A keynote panel on June 17th at VivaTech 2026 in Paris.
Am I the only one that thought RTK was Real-Time Kinematics used for precision with satellite navigation?
Need to be clear that "full http server in pure bash" is incorrect. Bash cannot listen on a TCP/UDP socket for incoming connections.
bash-web-server project builds a C language socket listener [0] that is dynamically loaded at run-time as a "built-in" module that makes the functionality available.
[0] https://github.com/bahamas10/bash-web-server/tree/main/loada...
$ systemctl cat public-inbox-watch@.timer
# /etc/systemd/system/public-inbox-watch@.timer
[Unit]
Description=Periodic fetch of public mailing list
[Timer]
# twice a day
OnCalendar=*-*-* 5,17:35
RandomizedDelaySec=1h
Persistent=true
[Install]
WantedBy=multi-user.targetThis should become a new round to compliment Mornington Crescent [0] on the BBC Radio programme I'm Sorry I Haven't A Clue [1]
[0] https://en.wikipedia.org/wiki/Mornington_Crescent_(game)
[1] https://en.wikipedia.org/wiki/I%27m_Sorry_I_Haven%27t_a_Clue
"So you can’t transparently bridge Ethernet devices’ MAC addresses through a WiFi client interface. This is why we need hostapd."
I think that is incorrect. hostapd handles the authentication side of things, but 4addr tuples are controlled by 'struct wireless_dev.use_4addr', and can be set by 'ip link set type bridge_slave ... proxy_arp_wifi on', `iw dev ... 4addr on', and if using systemd-networkd, with slave interface's
[Bridge]
ProxyARPWiFi=yes
Kernel then uses NL80211_IFTYPE_AP_VLAN and handles the proxy operation.
Barclays, with standard current accounts, provides several methods none of which are SMS. There's a separate pin-code device (called Pinsentry) that does TOTP and challenge-response, or passcodes for both telephone and Internet banking.
Firefox Nightly v149 has added experimental support via Settings > Firefox Labs:
Webpage Display
Media: JPEG XL
With this feature enabled, Nightly supports the JPEG XL (JXL) format. This is an enhanced image file format that supports lossless transition from traditional JPEG files. See bug 1539075 for more details."refused to load whatever distro I tried from SSD" sounds very much like a feature in AMI InsydeH2O firmware (and possibly others) where-by one has to manually "trust" the boot-loader file the boot menu entry points to. This doesn't seem to apply to Microsoft Windows boot loaders so I've always assumed the signing certificate is checked directly against the MS UEFI CA root rather than the intermediate 3rd party certificate that is used by Microsoft to sign distro shim files.
I have kept a screenshot of the firmware setup for years to remind me where the option can be found; looking at it now:
menu: Security > "Select UEFI file as trusted"
That would bring up a file-chooser where one can navigate the files in the EFI System Partition and select the distro's initial boot-loader file. For example, for a Debian install it would either or both of:
/EFI/debian/shimx64.efi /EFI/debian/grubx64.efi
Two devices I use - both running Debian, and both being open-source hardware to some degree or other:
PC Engines APU2, AMD x86_64, 4-core, 4GiB, 3x Gigabit Ethernet, 3 x mini PCIe, SIM slot, USB 3, Serial, SATA ports. Mine has dual band WiFi in one mPCIe, SSD in another.
Turris Mox, Marvel aarch64. This can expand via plug and go via a range of extension modules. I've got one with 25 Gigabit (3 x 8-port modules) Ethernet, 1 x SFP, 5 x USB3, Wifi, Serial.
This reminds me of a recent discussion about using a tarpit for A.I. and other scrapers. I've kept a tab alive with a reference to a neat tool and approach called Nepenthes that VERY SLOWLY drip feeds endless generated data into the connection. I've not had an opportunity to experiment with it as yet:
Have you checked out the The National Museum of Computing (TNMoC) archive. Last time I was there they had a rather good magazine collection going back to the early 1980s. It may be worth a call. I see they have an (incomplete) online catalogue:
When using LVM one can use the dm-integrity target to detect data corruption.
When using LVM there is no need to use separate mdadm (MD) based RAID - just use LVM's own RAID support.
I have a workstation with four storage devices; two 512GB SSDs, one 1GB SSD, and one 3TB HDD. I use LUKS/dm_crypt for Full Disk Encryption (FDE) of the OS and most data volumes but two of the SSDs and the volumes they hold are unencrypted. These are for caching or public and ephemeral data that can easily be replaced: source-code of public projects, build products, experimental and temporary OS/VM images, and the like.
dmsetup ls | wc -l
ls -1 /dev/mapper/${VG}-* | grep -E "${VG}-[^_]+$" | wc -l
This arrangement allows me to choose caching, raid, and any other device-mapper target combinations on a per-LV basis. I divide the file-system hierarchy into multiple mounted LVs and each is tailored to its usage, so I can choose both device-mapper options and file-system type. For example, /var/lib/machines/ is a LV with BTRFS to work with systemd-nspawn/machined so I have a base OS sub-volume and then various per-application snapshots based on it, whereas /home/ is RAID 1 mirror over multiple devices and /etc/ is also a RAID 1 mirror.
The RAID 1 mirrors can be easily backed-up to remote hosts using iSCSI block devices. Simply add the iSCSI volume to the mirror as an additional member, allow it to sync 100%, and then remove it from the mirror (one just needs to be aware of and minimising open files when doing so - syncing on start-up or shutdown when users are logged out is a useful strategy or from the startup or shutdown initrd).
Doing it this way rather than as file backups means in the event of disaster I can recover immediately on another PC simply by creating an LV RAID 1 with the iSCSI volume, adding local member volumes, letting the local volumes sync, then removing the iSCSI volume.
I initially allocate a minimum of space to each volume. If a volume gets close to capacity - or runs out - I simply do a live resize using e.g:
lvextend --resizefs --size +32G ${VG}/${LV}
lvextend --resizefs --size +32G ${VG}/${LV} ${PV}
btrfs filesystem resize max /srv/NAS/${VG}/${LV}To expand on this with an example. Adding a new device we'll call sdz to an existing Logical Volume Manager (LVM) Volume Group (VG) called "NAS" such that all the space on sdz is instantly available for adding to any Logical Volume (LV):
pvcreate /dev/sdz
vgextend NAS /dev/sdz
lvextend --size +128G --resizefs NAS/backup
"fsadm utility checks or resizes the filesystem on a device (can be also dm-crypt encrypted device). It tries to use the same API for ext2, ext3, ext4, ReiserFS and XFS filesystem."
If using BTRFS inside the LV, and the LV "backup" is mounted at /srv/backup, tell it to use the additional space using:
btrfs filesystem resize max /srv/backupI have a wake-alarm[0] that triggers 30 minutes before civil twilight, that is roughly 60 minutes before local sunrise.
In the northern hemisphere at 52 degrees it gets earlier by about 2 minutes each day (additional 4 minutes of daytime).
So I get more sleep and short days in winter and less sleep and longer days in summer. It's liberating basing schedule on it and not some arbitrary time.
[0] https://f-droid.org/packages/com.forrestguice.suntimeswidget...
For Starlink the User Terminal (antenna a.k.a. "Dishy") is a phased array. It tracks the satellite as it passes from west to east. Each satellite is in view for around 15 seconds - the phased array instantly flips from east to west and acquires the new in-view satellite in microseconds. There's no degradation in almost all 'flips' especially if the U.T. has an unobstructed view of the sky.
Yes, I use direct IPv6 peer-to-peer connections both outbound and inbound using the delegated prefix.
Even for a changing prefix, if operating a DNS authoritative server for a domain, any changes to the prefix can be quickly and automatically updated in both forward (AAAA) and reverse (PTR) resource records provided the TTL for those records is appropriately short, and thus allow almost seamless inbound via FQDNs. I do this with a bind9 (hidden) master locally that notifies external slave servers operated by a highly available, anycast, DNS service.
I've been using Starlink since early 2021 with IPv6 only internally. Starlink User Terminal hands out a /56 prefix (via DHCPv6) and mine has not changed in all that time so I wouldn't call it dynamic.
The User Terminal issues a router advertisement (RA) and my gateway gives itself an address in that /64 via SLAAC in addition to assigning itself an address from the /56 prefix.
If not using prefix delegation each host's address is dependent on their SLAAC policy - if not preferring stable addresses (e.g: EUI64) then of course the public address will vary (be dynamic) when using temporary "privacy" addresses.
My gateway delegates /60 sub-prefixes of the /56 and bare-metal hosts then either delegates /62 or advertises /64s from the /60 to VMs, containers, network namespaces and so forth.
As someone else described, I have my gateway also delegate ULA prefixes by changing just the first two octets of the public delegated prefix to fddc (fd = ULA, dc = "data center :) but otherwise identical and likewise on the bare-metal hosts, etc.
ULA is used for internal services; ISP delegated prefix for anything that needs public access.
Multicast-DNS takes care of internal hostnames; everything is ${hostname}.local
There's a separate VLAN for legacy IPv4-only devices that does NAT64 using a ULA prefix.
DNS64/NAT64 for the laggards like github.com that can't grok 128 bit addresses :)
The only time I have problems with web services is when their DNS advertises an AAAA resource record but their firewall/load-balancers/servers are not configured to allow/listen on it.
In my long experience of debugging and fixing ACPI errors exposed by Linux the reason MS Windows avoids (exposing to the operator) these firmware bugs is due to the fixes being incorporated into the Windows platform/chipset device drivers they ship.
I do something similar except that I do not allow wildcard reception - I create unique service-identifying user@ for each service I give an address to, and have a simple script that immediately adds that to the Postfix virtual table.
That way the SMTP server can reject all unknown user@ without accepting them in the first place - preventing spamming and some types of denial of service through resource starvation.
I also apply greylist based on a unique tuple (From, To, client IP address) so on first connection with that tuple valid SMTP clients need to re-deliver the email after a waiting period. Any subsequent delivers are accepted immediately.
This is often due to the total costs being externalised (pushed off to others) and therefore not reflecting the true cost of the replacement nor the costs of (safe) disposal of the old unit.
Externalised costs such as emissions from manufacturing of new raw materials (metals, plastics, gases, etc.), transportation, disposal, and more.
Obviously it depends on what exactly fails. I've kept 'white goods' going for over 20 years despite:
1) known defect where Hotpoint Fridge/Freezer evaporator thermistor fails due to freeze/defrost thermal cycle. Replaced more than 10 times; cost of new thermistor is pennies; time to replace (after initial explore) 10 minutes.
2) Freezer control PCB misreading thermistor; replace PCB: UK£35.
3) LG Washing machine bearing failures; replaced about 6 times; time to replace (after initial explore): 45 minutes.
Society as a whole in many countries is losing (or has already lost) the ability to be self-reliant and that lack makes people and communities generally more fragile.
Self-reliance is one of the drivers of hackers and tinkerers.
I wrote a Linux/BASH tester [0] for this recently that includes instructions on how to create simulated fake devices:
To create simulated (2GiB) fake devices
fallocate -l 1G fff_test.flash
DEV=$(losetup --show --find fff_test.flash); echo $DEV
DEVNUM=$(stat -c %Hr:%Lr $DEV); echo $DEVNUM
dmsetup create --concise "fff_wrap,,,,0 2097152 linear $DEVNUM 0, 2097152 2097152 linear $DEVNUM 0"
dmsetup create --concise "fff_drop,,,,0 2097152 linear $DEVNUM 0, 2097152 2097152 zero"
fake_flash_finder.bash /dev/mapper/fff_wrap
Capacity mismatch at LBA 2097152, data wrapped around to block 0. Size is most likely really 1073741824 bytes
fake_flash_finder.bash /dev/mapper/fff_drop
Capacity mismatch at LBA 2097152, data does not match what was written. Size is most likely really 1073741824 bytes
dd if=/dev/zero of=/dev/mapper/fff_wrap bs=64M status=progress conv=fdatasync
dd if=/dev/zero of=/dev/mapper/fff_drop bs=64M status=progress conv=fdatasync
dmsetup remove fff_wrap
dmsetup remove fff_drop
losetup --detach "$DEV"
rm fff_test.flash
I've still got the Notion Ink Adam with Pixel Qi transflective display and it beats everything modern in bright sunlight.
I wanted to check it out but they apparently can't even manage to do IPv6 correctly for their own service; advertising an AAAA record but not answering on it even for ICMPv6 let alone TCP 443/80 - apparently on AWS.
More important than that is text lends itself to searching for possibly obscure phrases to narrow down the possible candidates before even having to "consume" any information whereas with video that is challenging and very inefficient (time and energy-wise).
In Debian kernel we've very recently enabled building an ARM64 kernel flavour with 16KiB page size and we've discussed adding a 64KiB flavour at some point as is the case for PowerPC64 already.
This will likely reveal bugs that need fixing in some of the 70,000+ packages in the Debian archive.
That ARM64 16KiB page size is interesting in respect of the Apple M1 where Asahi [0] identified that the DART IOMMU has a minimum page size of 16KiB so using that page size as a minimum for everything is going to be more efficient.
[0] https://asahilinux.org/2021/10/progress-report-september-202...