https://time.com/6247678/openai-chatgpt-kenya-workers/
[src]
Alcuni estratti
To get those labels, OpenAI sent tens of thousands of snippets of text to an outsourcing firm in Kenya, beginning in November 2021. Much of that text appeared to have been pulled from the darkest recesses of the internet. Some of it described situations in graphic detail like child sexual abuse, bestiality, murder, suicide, torture, self harm, and incest.
One Sama worker tasked with reading and labeling text for OpenAI told TIME he suffered from recurring visions after reading a graphic denoscription of a man having sex with a dog in the presence of a young child. “That was torture,” he said. “You will read a number of statements like that all through the week. By the time it gets to Friday, you are disturbed from thinking through that picture.” The work’s traumatic nature eventually led Sama to cancel all its work for OpenAI in February 2022, eight months earlier than planned.
In February, according to one billing document reviewed by TIME, Sama delivered OpenAI a sample batch of 1,400 images. Some of those images were categorized as “C4”—OpenAI’s internal label denoting child sexual abuse—according to the document. Also included in the batch were “C3” images (including bestiality, rape, and sexual slavery,) and “V3” images depicting graphic detail of death, violence or serious physical injury, according to the billing document. OpenAI paid Sama a total of $787.50 for collecting the images, the document shows.
But the need for humans to label data for AI systems remains, at least for now. “They’re impressive, but ChatGPT and other generative models are not magic – they rely on massive supply chains of human labor and scraped data, much of which is unattributed and used without consent,” Andrew Strait, an AI ethicist, recently wrote on Twitter. “These are serious, foundational problems that I do not see OpenAI addressing.”
[src]
Alcuni estratti
To get those labels, OpenAI sent tens of thousands of snippets of text to an outsourcing firm in Kenya, beginning in November 2021. Much of that text appeared to have been pulled from the darkest recesses of the internet. Some of it described situations in graphic detail like child sexual abuse, bestiality, murder, suicide, torture, self harm, and incest.
One Sama worker tasked with reading and labeling text for OpenAI told TIME he suffered from recurring visions after reading a graphic denoscription of a man having sex with a dog in the presence of a young child. “That was torture,” he said. “You will read a number of statements like that all through the week. By the time it gets to Friday, you are disturbed from thinking through that picture.” The work’s traumatic nature eventually led Sama to cancel all its work for OpenAI in February 2022, eight months earlier than planned.
In February, according to one billing document reviewed by TIME, Sama delivered OpenAI a sample batch of 1,400 images. Some of those images were categorized as “C4”—OpenAI’s internal label denoting child sexual abuse—according to the document. Also included in the batch were “C3” images (including bestiality, rape, and sexual slavery,) and “V3” images depicting graphic detail of death, violence or serious physical injury, according to the billing document. OpenAI paid Sama a total of $787.50 for collecting the images, the document shows.
But the need for humans to label data for AI systems remains, at least for now. “They’re impressive, but ChatGPT and other generative models are not magic – they rely on massive supply chains of human labor and scraped data, much of which is unattributed and used without consent,” Andrew Strait, an AI ethicist, recently wrote on Twitter. “These are serious, foundational problems that I do not see OpenAI addressing.”
TIME
Exclusive: OpenAI Used Kenyan Workers on Less Than $2 Per Hour to Make ChatGPT Less Toxic
A TIME investigation reveals the difficult conditions faced by the workers who made ChatGPT possible
Nuove osservazioni
Con il seguente setup:
• driver
(niente modesetting)
• kernel mode setting attivo
(è una cosa diversa)
• accelerazione
• opzione
• nessun compositor
= le applicazioni opengl soffrono di gravi cali ogni ~3s di durata ~½s
Se in aggiunta viene forzata l'opzione lato MESA
Usando
Usando
Utilità
Con il seguente setup:
• driver
intel su xorg(niente modesetting)
• kernel mode setting attivo
(è una cosa diversa)
• accelerazione
SNA• opzione
TearFree attiva• nessun compositor
= le applicazioni opengl soffrono di gravi cali ogni ~3s di durata ~½s
Se in aggiunta viene forzata l'opzione lato MESA
vblank_mode=3, questo problema scompareUsando
intel + l'accelerazione UXA, firefox non avrà mai vsync, nonostante tutte le altre opzioni.Usando
modesetting non c'è vsync di default, e aggiungendo SNA (il default è glamor) và in software rasterUtilità
export LIBGL_DEBUG=verbose
less /var/log/Xorg.0.logMedia is too big
VIEW IN TELEGRAM
My firefox experience
Il video è a 60fps per evidenziare i brevi stutter nonostante il vsync
Solo firefox ha questo problema
Il video è a 60fps per evidenziare i brevi stutter nonostante il vsync
Solo firefox ha questo problema
Sad, ovviamente ho provato anche con
sudo
$ su -c 'whoami; schedtool -F -p 40 2475 2268 2267'
Password:
root
ERROR: could not set PID 2475 to F: SCHED_FIFO - Operation not permitted
ERROR: could not set PID 2268 to F: SCHED_FIFO - Operation not permitted
ERROR: could not set PID 2267 to F: SCHED_FIFO - Operation not permittedUn canale Telegram potrà non essere il posto ideale per chiederlo [ancor meno questo], tuttavia-
Come state?
Come state?
Il kernel Linux Clear non ha ntfs3 !??
$ udisksctl mount --block-device /dev/sdd1
Error mounting /dev/sdd1: GDBus.Error:org.freedesktop.UDisks2.Error.Failed: Error mounting /dev/sdd1 at /run/media/micro/#####: Filesystem type ntfs3 not configured in kernel.Fantastico, non sono il solo ad aver avuto problemi con ntfs3
Torno a ntfs-3g definitivamente
https://old.reddit.com/r/linux/comments/uca3fu/ntfs3_driver_is_orphan_already_what_we_do/i69453f/
Torno a ntfs-3g definitivamente
https://old.reddit.com/r/linux/comments/uca3fu/ntfs3_driver_is_orphan_already_what_we_do/i69453f/
journalctl -u micro
Fantastico, non sono il solo ad aver avuto problemi con ntfs3 Torno a ntfs-3g definitivamente https://old.reddit.com/r/linux/comments/uca3fu/ntfs3_driver_is_orphan_already_what_we_do/i69453f/
Relative configurazioni
> /etc/fstab
> /etc/fstab
(ntfs3 -> ntfs)> /etc/udev/rules.d/80-ntfs3.rules
#SUBSYSTEM=="block", ENV{ID_FS_TYPE}=="ntfs", ENV{ID_FS_TYPE}="ntfs3"
> /etc/udisks2/mount_options.conf[defaults]
ntfs_defaults=uid=$UID,gid=$GID,relatime,windows_names,big_writes
ntfs_allow=uid=$UID,gid=$GID,umask,dmask,fmask,locale,norecover,ignore_case,windows_names,compression,nocompression,big_writes
ntfs3_defaults=uid=$UID,gid=$GID,relatime,acl,prealloc
ntfs3_allow=uid=$UID,gid=$GID,umask,dmask,fmask,nls,nohidden,sys_immutable,discard,force,sparse,showmeta,prealloc,no_acs_rules,acl
btrfs_defaults=compress=zstd:3,autodefrag
btrfs_allow=compress,compress-force,datacow,nodatacow,datasum,nodatasum,autodefrag,noautodefrag,degraded,device,discard,nodiscard,subvol,subvolid,space_cache
Pandoc recursive batch convert docx2pdf
margins='left=2.5cm,right=2.5cm,top=1.6cm,bottom=2cm'
while read -r d; do
echo "- $d"
noext="${d%.*}"
out="$noext.pdf"
[[ -e "$out" ]] && continue
pandoc --pdf-engine=lualatex \
-V 'mainfont:carlito' \
-V 'papersize:a4' \
-V "geometry:$margins" \
-o "$out" "$d"
done < <(find . -type f -iname '*.docx')Questo invece ha messo degli screen generati da AI
https://play.google.com/store/apps/details?id=com.virus.remover
https://play.google.com/store/apps/details?id=com.virus.remover
journalctl -u micro
https://tex.stackexchange.com/questions/503/why-is-preferable-to
https://tex.stackexchange.com/questions/173717/is-it-ever-bad-to-use-left-and-right
Latex ti fa pensare al contenuto e non alla forma
Latex ti fa pensare al contenuto e non alla forma
Latex ti fa pensare al contenuto e non alla forma
Latex ti fa pensare al contenuto e non alla forma
Latex ti fa pensare al contenuto e non alla forma
Latex ti fa pensare al contenuto e non alla forma
Latex ti fa pensare al contenuto e non alla forma
Latex ti fa pensare al contenuto e non alla forma