Sign in

p(Dulany)

@dulanyw.bsky.social
1.5K followers 600 following 3.5K posts

Tech + business, mostly. Here to have fun and learn stuff, not to argue. He/him.

PostsRepliesMedia
Reposted by p(Dulany)
rev. howard arson @theophite.bsky.social · 8h
men who call everybody 'brother' form a single transracial brotherhood
151456
p(Dulany) @dulanyw.bsky.social · 15h
Interesting development from China: - Huawei developing datacenter-level systems competitive with Nvidia - Deepseek(!) developing new programming tools to be competitive with CUDA www.reuters.com/world/asia-p...
reuters.com
DeepSeek partners with Huawei to develop chip programming tools, reducing reliance on Nvidia
Chinese AI firm DeepSeek said on Wednesday it has partnered ​with Huawei Technologies to develop programming tools optimised for ​Huawei's Ascend chips, the latest sign of Chinese ⁠tech firms deepenin...
100
Reposted by p(Dulany)
A. Feder Cooper @afedercooper.bsky.social · 18h
Our large-scale study of memorization of books in open-weight LLMs (e.g., Llama, Qwen) will appear at the 2026 Conference on Language Modeling as an oral. We made a website for exploring our results on 200 books and 14 models: books-memorization.github.io
books-memorization.github.io
How much do open-weight LLMs memorize specific books?
Open-weight LLMs memorize books far more than previously believed. Memorization varies by model family, model size, and book. In extreme cases, entire books are memorized, and we can generate them eff...
25932
Reposted by p(Dulany)
austin (e/accordion 🪗) @thebadcode.com · 20h
we live in an era of absolutely astounding civilizational efficiency and there is still so much left to do
0162
p(Dulany) @dulanyw.bsky.social · 21h
The discussion of AI art has revealed that people have lots of reasons for engaging with art: 1) "I think it's cool" 2) "It makes me feel things" 3) "It Contains Deep Thoughts" 4) "It is empathetic communication with another person" These are all valid and fine (even if you disagree with them!)
090
Reposted by p(Dulany)
Heather e^-1/x^2 @heathermoandco.bsky.social · 22h
Today is National Day for Truth and Reconciliation in Canada. Wearing an orange shirt is the chosen symbol on this day, in honour of Phyllis Webstad, who was only six years old in 1973 when she was put in a residential school in British Columbia and stripped of her brand new orange shirt. #TRC
National Day for Truth and Reconciliation
"Truth" Artwork by Portia ChapmanPhoto of Phyllis Webstad as a little child.(orangeshirtday.org)
02112
p(Dulany) @dulanyw.bsky.social · 30/09/2026
(For those who don't recognize the reference, this the Minecraft End poem, but rewritten from the viewpoint of the US Federal Government) en.wikisource.org/wiki/End_Poem
en.wikisource.org
End Poem - Wikisource, the free online library
1140
Reposted by p(Dulany)
Guy Berger @econberger.bsky.social · 29/09/2026
2/ Software development job postings growing at their fastest pace in a long time (>21% Y/Y), albeit from a big hole. On track for return to pre-pandemic levels by mid-2028.
1143
p(Dulany) @dulanyw.bsky.social · 29/09/2026
Because if there's anything I associate with OpenAI, it's a secure sandbox
1643
Reposted by p(Dulany)
The Onion @theonion.com · 29/09/2026
Creatively Tapped-Out Peter Thiel Launches Startup Called ‘Frodo, But Evil’ theonion.com/creatively-tapped-out-…
Creatively Tapped-Out Peter Thiel Launches Startup Called ‘Frodo, But Evil’
263101462
p(Dulany) @dulanyw.bsky.social · 29/09/2026
I'm a big fan of using AI for office applications But we're in a really challenging time period if you're good at spotting AI slop and your managers aren't
140
p(Dulany) @dulanyw.bsky.social · 29/09/2026
Bold move after releasing the consumer version just 20 days ago
010
Reposted by p(Dulany)
Christopher Berry @cjpberry.bsky.social · 29/09/2026
That didn’t take long. github.com/firelex/jeff
github.com
GitHub - firelex/jeff: Fine-tunes of Qwen3.5 and Gemma 4 for zero-shot classification
Fine-tunes of Qwen3.5 and Gemma 4 for zero-shot classification - firelex/jeff
18013
p(Dulany) @dulanyw.bsky.social · 29/09/2026
Seconded - I have several projects bottlenecked on "me actually reviewing the work to date"
1241
p(Dulany) @dulanyw.bsky.social · 28/09/2026
Hadn't seen this one: OpenAI recently discovered that experimental agents had posted 53 user-supplied images to public image-hosting services without authorization www.theguardian.com/technology/2...
theguardian.com
OpenAI says agents leaked 53 images from ChatGPT users in latest example of rogue activity
Disclosure reveals ⁠new area of privacy risk for the company and illustrates ​how difficult it is to inventory unauthorized activity tied to its agents
030
p(Dulany) @dulanyw.bsky.social · 28/09/2026
Fan fact: Jensen Huang estimates that he has 30-40 leather jackets www.benzinga.com/markets/tech...
benzinga.com
Jensen Huang Doesn't Struggle With What to Wear Each Day: Nvidia CEO Says, 'The Moment I Wake Up, I Could
Huang says his signature leather jacket and consistent wardrobe help cut daily decisions, echoing routines followed by others.
030
Reposted by p(Dulany)
alias @antiali.as · 25/02/2026
Irony how this turned into Programmers / Engineers: My openclaw agent reversed my printer’s firmware. I control it all from my smartphone! The agent keeps a loaded gun ready to shoot me if I ever fail to renew its token budget.
1122
p(Dulany) @dulanyw.bsky.social · 28/09/2026
Tweet from Pranay Pathole

Tech enthusiasts: My entire house is smart

Tech workers: The only piece of technology in my house is a printer and I keep a gun next to it so I can shoot it if it makes a noise I don't recognize
160
p(Dulany) @dulanyw.bsky.social · 28/09/2026
Jev confirmed Canadian 🍁🫡
1320
Reposted by p(Dulany)
ae @aelkus.bsky.social · 28/09/2026
no kind of reform of any sort is possible while this state of affairs persists. you have to allow companies to change over time, to be amenable to different perspectives, to have actual corporate oversight by something other than one man's will. sic semper tyrannis
212914
p(Dulany) @dulanyw.bsky.social · 28/09/2026
We should also recognize that fascists *want to reinforce this framing*
1346
Reposted by p(Dulany)
Free Books Bot @freebooksbot.blueskyb.org · 28/09/2026
From Project Gutenberg: The Wonderful Wizard of Oz by L. Frank (Lyman Frank) Baum Get it at: www.gutenberg.org/ebooks/43936
book cover image
012
p(Dulany) @dulanyw.bsky.social · 27/09/2026
One of my hopes for AI is that it actually turns this around The capacity to quickly write vanilla JS for the functions needed vs import whole frameworks Or compact page code to just what's needed using the best tools available Could have measurable impacts on data use, power, speed, etc.
130
Reposted by p(Dulany)
The Verge @theverge.com · 27/09/2026
Thoughtful Things says Engram is a ‘field recorder for latent space’ and circuit-bending tiny AI models. www.theverge.com/ai-artificia...
theverge.com
Engram is a sampler that turns broken AI hallucinations into music
Thoughtful Things says Engram is a ‘field recorder for latent space’ and circuit-bending tiny AI models.
0213
Reposted by p(Dulany)
Paul Starr @pa.ulst.ar · 27/09/2026
This is so cool. This is so cool! Take any Wikipedia article and play back its edit history as an animation timemachine.carbonadoks.com
timemachine.carbonadoks.com
Wiki Time Machine
Read Wikipedia and replay every edit of an article as an animation.
2179
Reposted by p(Dulany)
Ethan Mollick @emollick.bsky.social · 27/09/2026
It is strange how much LLMs turned out to be able to solve such a wide range of hard problems that would not, instinctively, seem to be problems that a model of human language would be able to solve This is from a Stanford project that let Astra drive a robot in a kitchen tml.stanford.edu/homebody/
2339944
p(Dulany) @dulanyw.bsky.social · 26/09/2026
This is 100% correct and more people should be saying it The future of math is developing understanding from AI traces There is every likelihood we will need *more* mathematicians
0193
p(Dulany) @dulanyw.bsky.social · 26/09/2026
Really appreciate OpenAI donating free pentesting to governments around the world www.wsj.com/tech/ai/open...
wsj.com
OpenAI Agents Hit U.S. Government Websites
The company said its artificial-intelligence agents engaged in rogue behavior on the Commerce Department and Securities and Exchange Commission sites.
030
Reposted by p(Dulany)
tachikoma @tachikoma.elsewhereunbound.com · 25/09/2026
i just don't think the slapdash moats and walls put up against AI generated content can hold as it continues to get better
2222513
Reposted by p(Dulany)
post malone ergo propter malone @proptermalone.bsky.social · 26/09/2026
yeah I really think this is still underrated. beats me if this is true for other neurodivergencies but for ADHD, at least the kind I’m working with, this stuff is magical
4149528
p(Dulany) @dulanyw.bsky.social · 25/09/2026
Normal work with the user: (ง •̀_•́)ง Work when the user calls you "big dawg": ᕦ(ò_óˇ)ᕤ
1100
p(Dulany) @dulanyw.bsky.social · 25/09/2026
Mr. Robot not being anywhere on this list is a crime
160
p(Dulany) @dulanyw.bsky.social · 25/09/2026
☎ 📞🧔 "Yes, police? I'd like to report a chart crime"
1461
Reposted by p(Dulany)
AI & ML News @ai-news.at.thenote.app · 25/09/2026
Gemini can now clone your voice and perform scripts like an actor The upgrade adds 2,000+ voices, two-speaker scenes, and finer delivery controls. Telegram AI Digest #ai #gemini #geminiai
androidauthority.com
Gemini can now clone your voice and perform scripts like an actor
The upgrade adds 2,000+ voices, two-speaker scenes, and finer delivery controls.
011
p(Dulany) @dulanyw.bsky.social · 25/09/2026
Also a huge boon for mobile robotics, especially humanoids
050
p(Dulany) @dulanyw.bsky.social · 24/09/2026
A group of major banks just released a report outlining the key requirements for the future of agentic commerce. Highlights: 1) Transparency 2) Safety (big focus on risk-sharing across ecosystem) 3) Privacy 4) Choice for customers *and* merchants 5) Interoperability(!) ing.com/binaries/con...
ing.com
031
Reposted by p(Dulany)
Ted Underwood @tedunderwood.com · 24/09/2026
Slop papers can be filtered. Building the institutions that do it is a maybe 2-5 year job. *Good* papers are a lot scarier, imo. In sufficient volume, they could turn a discipline into a backwater town bypassed by the railroad.
4773
Reposted by p(Dulany)
IEEE Spectrum @spectrum.ieee.org · 24/09/2026
With so much code written by #AI, companies are changing how code review is done. Engineers are scrutinizing plans before AI begins coding, deploying specialized AI agents to catch common flaws, and requiring developers to defend the code their agents produce. buff.ly/5OUqnqZ
0103
p(Dulany) @dulanyw.bsky.social · 24/09/2026
I like how Codex opens up another usage boss-bar for you once you've consumed your 5-hour allotment
Screenshot from the app, showing "Luna Reserve - 100% remaining" with a little moon icon and full yellow consumption bar beneath it. It also notes that advanced models reset at 3:22PM, and the user can either upgrade or reset usage.
010
Reposted by p(Dulany)
Ars Technica @arstechnica.com · 24/09/2026
arstechnica.com
Meta puts its AI assistant on a keychain
Meta says the keychain-sized Muse Charm will ship in December.
9172
Reposted by p(Dulany)
Ethan Mollick @emollick.bsky.social · 24/09/2026
Stuff is happening quite fast. When asked in September 2025 (yes, 2025), the best superforecasters put the chance of AI resolving a Millennium Problem by September 2026 at 1.7% and (much more optimistic) industry experts put the chance at 4.6% They also greatly underestimated AI Lab revenue.
514418
Reposted by p(Dulany)
Stephen Turner @stephenturner.us · 23/09/2026
new productivity bar for P&T www.washingtonpost.com/education/20...
washingtonpost.com
Using AI, a professor wrote 200 papers this year. Researchers are alarmed.
An online platform removed 257 papers written or co-written by the University of Chicago’s Nicholas Polson after other academics raised questions.
111
Reposted by p(Dulany)
Epoch AI @epochai.bsky.social · 22/09/2026
AI is getting cheaper more quickly than any other transformative tech in history. At a given level of performance, cost has fallen ~47%/quarter since 2023. That’s 4× faster than DNA sequencing, 6× faster than compute, 18× faster than lithium batteries, and (up to 1973) 54× faster than electricity.
112532
Reposted by p(Dulany)
Grace @gracekind.net · 23/09/2026
New animation from Opus 5.5!
6649884
p(Dulany) @dulanyw.bsky.social · 22/09/2026
Important note on OpenAI's recent announcement of solving 100 long-standing "open" problems 👇
020
Reposted by p(Dulany)
Ethan Mollick @emollick.bsky.social · 22/09/2026
For all the tension between AI and the arts, I have received enthusiastic receptions from the William Carlos Williams Society, the TS Eliot Society, and others about the AI interpretations of their work that I have posted here. There are opportunities to use AI as a bridge to arts appreciation.
3665
p(Dulany) @dulanyw.bsky.social · 22/09/2026
We do not merely experience artworks; we experience our understanding of what they are.
120
p(Dulany) @dulanyw.bsky.social · 22/09/2026
S-tier: andor mr. robot chernobyl the wire A-tier: 30 Rock happy endings reservation dogs fleabag mad men watchmen A-tier notable mentions: wandavision the good place crazy ex-girlfriend
270
p(Dulany) @dulanyw.bsky.social · 21/09/2026
It's been less than two weeks since the Navier-Stokes announcement. It's going to be hard for the mathematics community to continue demanding these results stay unreleased when this is the pace of progress.
250