#technology Timeline

239d

Our commitment to Windows quality

1y

AI slop is threatening journalism by churning out clickbaity versions of original stories, capturing web traffic, and destroying readers' trust in media. NiemanLab's Ben Paviour takes a look at how it's proliferating, with the help of Google.

https://flip.it/6VypfN

#ArtificialIntelligence #AI #Slop #Journalism #Media #MediaIndustry #Google #Technology #Tech

0 0 1 View Post & Replies See Original

240d

> For example, Google reduced our headline “I used the ‘cheat on everything’ AI tool and it didn’t help me cheat on anything” to just five words: “‘Cheat on everything’ AI tool.” It almost sounds like we’re endorsing a product we do not recommend at all.

https://www.theverge.com/tech/896490/google-replace-news-headlines-in-search-canary-coal-mine-experiment

#news #TechNews #technology #google #search #AI #LLMs #enshittification

Edited 240d ago

1 0 1 View Post & Replies See Original

1y

#lispyGopherClimate #live 000UTC (Tuesday night in Americas various) #technology #podcast

https://anonradio.net:8443/anonradio

A quiet episode.

#climateCrisis #haiku from @kentpitman
#knowledgePreservation
- and the climate ^
- Scientific computing, with style notes from @AmenZwa again

Revisiting Dungeon Crawler Carl, with descriptions (= light spoilers) and the relation of games and specialised knowledge

Access to knowledge, #lisp, #eev, nonmonotonic logic, testing, proofs and the delete problem.

The gopher and lisp alien wade through flooded wreckage flanked by two demons carrying radio equipment

LISPY GOPHER SHOW
every 000UTC Wednesday on anonradio.net

a unix_surrealism banner

ALT

0 0 0 View Post & Replies See Original

1y

DeepSeek releases DeepSeek OCR

LLMs totally choke on long context because of that O(n2) scaling nightmare. It’s the core scaling problem for almost all modern LLMs because of their self-attention mechanism.

In simple terms, for every single token in the input, the attention mechanism has to look at and calculate a score against every other single token in that same input.

So, if you have a sequence with n tokens, the first token compares itself to all n tokens. The second token also compares itself to all n tokens… and so on. This means you end up doing n*n, or n^2, calculations.

This is a nightmare because the cost doesn’t grow nicely. If you double your context length, you’re not doing 2x the work; you’re doing 2^2=4x the work. If you 10x the context, you’re doing 10^2=100x the work. This explodes the amount of computation and, more importantly, the GPU memory needed to store all those scores. This is the fundamental bottleneck that stops you from just feeding a whole book into a model.

Well, DeepSeek came up with a novel solution to just stop feeding the model text tokens. Instead, you render the text as an image and feed the model the picture. It sounds wild, but the whole point is that a huge wall of text can be “optically compressed” into way, way fewer vision tokens.

To do this, they built a new thing called DeepEncoder. It’s a clever stack that uses a SAM-base for local perception, then a 16x convolutional compressor to just crush the token count, and then a CLIP model to get the global meaning. This whole pipeline means it can handle high-res images without the GPU just melting from memory activation.

And the results are pretty insane. At a 10x compression ratio, the model can look at the image and “decompress” the original text with about 97% precision. It still gets 60% accuracy even at a crazy 20x compression. As a bonus, this thing is now a SOTA OCR model. It beats other models like MinerU2.0 while using fewer than 800 tokens when the other guy needs almost 7,000. It can also parse charts into HTML, read chemical formulas, and understands like 100 languages.

The real kicker is what this means for the future. The authors are basically proposing this as an LLM forgetting mechanism. You could have a super long chat where the recent messages are crystal clear, but older messages get rendered into blurrier, lower-token images. It’s a path to unlimited context by letting the model’s memory fade, just like a human’s.

DeepSeek-OCR

1 1 0 View Post & Replies See Original

310d

With so much of the internet now being AI slop, a "smiley face with squinting eyes" emoji sure will come in handy!

https://blog.emojipedia.org/draft-emoji-list-for-2026-2027/

#news #technology #emoji

0 0 1 View Post & Replies See Original

303d

What's New for Mozilla Firefox Release 147.0

Examine that latest new features, improvements and fixes for Mozilla Firefox version 147.0 broken into easy to follow headings.

https://www.adamsdesk.com/posts/mozilla-firefox-whats-new-147/

#blog #tech #technology #news #Firefox #Mozilla

ALT

0 0 0 View Post & Replies See Original

310d

Microsoft CEO Begs Users to Stop Calling AI Content "Slop"

microsoft-satya-nadella-ai-slop

1 0 0 View Post & Replies See Original

310d

Exquisite.social is a general Glitch Mastodon server with a slight focus on technology, privacy and *BSD. We are here to play and make friends in a supportive atmosphere.

This server has a post size of up to 2500 characters.

https://exquisite.social

You can find out more at https://exquisite.social/about or contact the admin account @exquisite

#FeaturedServer #Technology #Privacy #BSD #GlitchMastodon #Mastodon #Fediverse #FreeFediverse

0 0 1 View Post & Replies See Original

309d

“I feel like we're rebooting to the last version of the Internet that I felt uncomplicated joy about.”

@Casey, @kevin and @pj on why they set up their Forkiverse fediverse server.

https://www.searchengine.show/the-fediverse-experiment/

#Fediverse #Federation #ActivityPub #Forkiverse #OpenSocial #OpenSocialWeb #Internet #Technology #Tech

0 0 1 View Post & Replies See Original

1y

Nike says its first ‘powered footwear’ is like an e-bike for your feet

https://www.theverge.com/news/805664/nike-powered-show-project-amplify

#technology #press

0 0 1 View Post & Replies See Original

111d

Bonus video this week!

I read the story of Samuel Tunick, who has a criminal case pending against him for using a 'duress password' to delete his date during a US border entry, and couldn't resist making a video.

If you enjoy the video, please please please tap the like button. This helps us smaller creators immensely.

#Tech #Technology #GrapheneOS #Video #SamuelTunick #YouTube #Privacy #Fascism

https://youtu.be/YSvDehDB8Kc

0 0 0 View Post & Replies See Original

1y

Batch Scan Photos With SANE's Scanimage On GNU/Linux Command Line

Turn a painful, time consuming and repetitive task of batch scanning photographs into a joyful one using SANE’s Scanimage on a GNU/Linux command line.

https://www.adamsdesk.com/posts/batch-scan-photo-sane-scanimage/

#blog #commandline #linux #tech #technology

A bold title reads 'Batch Scan Photos' next to a flatbed image scanner with a cartoon styled family Polaroid portrait placed onto the bed.

ALT

0 0 0 View Post & Replies See Original

77d

#lispyGopherClimate very haphazard Sunday-morning-in-Europe #peertube #live weekly #technology #podcast

@alexshendi #lisp and possibly @jhlagado we are going to meet up in #lambdaMOO and on the livestream and talk about lisp and retrocomputing and modern computing.

https://toobnix.org/w/g2AftRUaBvEVGfJYMY1MHr

John Hardy's git feat. vintage game tape WAV https://github.com/jhlagado

and @ramin_hal9001 ! And @rat in-MOO!

Computer origins, #xlisp, #z80, #formalMethods ..! #bookstodon #neuromancer #godelEscherBAch

Lispy gopher climate get Sandewalled banner ! Actually 8UTC toobnix.org.

ALT

Edited 77d ago

1 0 0 View Post & Replies See Original

234d

Am I a Tech Bro? 18 statements. Agree or Disagree. Take the quiz: https://amiatechbro.com

I think 0% was not bad.

#BigTech #AI #Technology

ALT

1 0 0 View Post & Replies See Original

123d

Chinese memory maker CXMT is expanding so fast it could hit 350,000 DRAM wafers per month by 2026. That puts it right behind Micron at 375,000. Micron has decades of history while CXMT just showed up to serve China. Insane how quickly they caught up.

https://www.techpowerup.com/350726/chinese-cxmt-to-match-microns-dram-manufacturing-capacity-this-year

#china #technology #news

0 0 0 View Post & Replies See Original

112d

#lispyGopherClimate Sunday-morning-in-Europe #live #technology #lisp #podcast every week since 2022. #ARCHIVED

https://toobnix.org/w/pxKfPQut7XZVpuPP3pGujH

Floating point numbers in ANSI CL very lightly and what people are talking about on irc.

@cwebber 's vibrating bobsled #ai article, #emacs and a Canadian sociologist falling for her "research ai boyfriend"
https://dustycloud.org/blog/faulty-towers-vibe-sickness-and-the-vibe-bobsled/
https://www.cbc.ca/news/canada/kitchener-waterloo/whats-it-like-having-an-ai-boyfriend-2-waterloo-researchers-decided-to-find-out-9.7281399

Talking about the state of #zine issue 0 (released tuesday-in-america no-matter-what) #peertube #fedi

A very spooky green lisp alien lispy gopher show banner, with the gopher looking on from behind.

ALT

Edited 112d ago

1 0 0 View Post & Replies See Original

81d

#lispyGopherClimate #live NOW at 0UTC Wednesday "Tuesday night in the Americas", every week since 2022 #lisp #technology #climateCrisis #podcast !

https://toobnix.org/w/kSbsFxarX8FAb8c55BsZVt

Join in live on #lambdaMOO . https://lambda.moo.mud.org/ !

- It is yduJ's halfbirthday @nosrednayduj https://hachyderm.io/@nosrednayduj/117153529001379734

- #haiku by @kentpitman

Pursuant to great moomail I got from slak, I want to talk about postscarcity lambdamoo virtual reality, UBI, and #VR and #lisp.

+ being impactful.

@ramin_hal9001

Lispy gopher climate unix_surrealism LISPY GOPHER SHOW banner featuring the gopher in interlisp shirt and lisp cthuluian alien. Sunday 8UTC toobnix.org, Wednesday 0UTC anonradio.net

ALT

Edited 81d ago

1 0 0 View Post & Replies See Original

109d

When Microsoft reviewed the early results of Anthropic’s new AI model for cybersecurity, it was a sobering scene: Hundreds of bugs were identified, many of them critical.

What followed was a “mad dash” to patch them — but many more remain.

https://www.propublica.org/article/anthropic-mythos-microsoft-software-vulnerabilities?utm_source=mastodon&utm_medium=social&utm_campaign=mastodon-post

#News #Cybersecurity #Tech #Microsoft #Technology #Anthropic #AI #ArtificialIntelligence #Hacking

3 0 0 View Post & Replies See Original

297d

Google’s New Android Update Locks Your Apps And Hides Your Messages

Edited 297d ago

1 1 0 View Post & Replies See Original

297d

"But, as Michèle Martin documents in her paper “The Rulers of the Wires”, middle-class women’s “‘delinquent’ telephone activities” took the Telegram Men by surprise. This delinquency included not just social telephone calls in the afternoon, but planned and impromptu meet-ups on party lines where isolated women could dial in and get an update from their community.

This was damned at the time by the Telegram Men as futile and frivolous, but that didn’t stop it from happening."

https://medium.com/careful-industries/delinquent-telephone-activity-f75f815d6e9a

#essay #community #technology

0 0 1 View Post & Replies See Original

305d

Owners, not renters: Mozilla's open source AI strategy

0 1 0 View Post & Replies See Original

108d

"LinkedIn, a social network awash with long AI-generated posts from executives and other corporate workers, has introduced a new button that users can click to flag if a post “seems like AI slop,” according to [@404mediaco's] own tests."

https://www.404media.co/linkedin-introduces-a-seems-like-ai-slop-button/

#news #technology #linkedin #AI #LLMs #slop #enshittification