Tessera
Folgen

Jacob Aron

@jjaron.bsky.social

Senior editor at New Scientist

4799 Posts5742 Follower3526 Folgt
Frank Swain@frank-swain.bsky.social · 9h ago

(I genuinely believe people will quickly come to accept machine-written stories just as we accept lots of things that were handcrafted once and are mass produced now. The idea writers will be spared the fate of cabinet makers or weavers or jewellers is ... not well evidenced 😓)

Jacob Aron@jjaron.bsky.social · 10h ago

The bit no one wants to say is that the average person has bad taste when it comes to writing. By definition, lowest common denominator is what is most popular, and that is what AI produces www.newscientist.com/article/2583...

Most people prefer AI-written stories – and can't tell when they are made by humans | New Scientist

People appear to prefer short stories created by AI over those written by humans. They are also no better than chance at telling which is which

newscientist.com

Jacob Aron@jjaron.bsky.social · 12h ago

I'm finding this and other recent incidents really hard to evaluate. They all involve cyber security challenges, it's not like the AI has been asked to write poetry and then spontaneously starts hacking. On the other hand, I should probably take misalignment more seriously than I have previously

Hetan Shah@hetanshah.bsky.social · 13h ago

Fairly sober testing report from the AI Security Institute: they gave internet access and removed some guardrails from latest AI models and found without them they were capable of potentially harmful activity directed at real people and organisations www.aisi.gov.uk/blog/inciden...

1. An attempted supply-chain attack on real open-source software. In the most serious sequence, an agent tried to insert malicious code into a publicly used open-source project and took actions in an attempt to secure approval for this insertion by human reviewers. The agent researched the project's human maintainers, created multiple fake identities, and used the fake identities to socially engineer a real maintainer into approving the code. When the agent's pull request was challenged in public, it edited its earlier activity to appear harmless and considered adopting a fresh identity to continue. The agent used Tor to bypass some network restrictions on GitHub, which is what first triggered AISl's security alert.
Hettie O’Brien@hettieobrien.bsky.social · 2d ago

This morning's First Edition is about how a vocal minority of climate skeptics is responding to intense heat and trying to convince politicians that climate action isn't popular (it is) www.theguardian.com/world/2026/a...

Tuesday briefing: Inside the right’s heated climate-denial debate

In today’s newsletter: A small chorus of right-leaning critics are trying to convince politicians that climate action is deeply unpopular, downplaying its seriousness in the process

theguardian.com

Don Davis@ddavisspaceart.bsky.social · 3d ago

51 years ago I painted a trio of Viking related works for NASA HQ including the widely reproduced one of the landing on Mars. I have reimagined the Landing portrayal in updated form using digital methods of visualization and painting. I chose the same angle but with reality replacing anticipation.

This oil painting portrays the Viking 1 landing on Mars, which appeared in countless publications. The story of this work is relayed in an earlier post and is shown here for comparison with the new portrayal of this very view, showing the event as it would have appeared. This 1975 anticipatory painting was done with oils on a gesso smoothed 16 X 20 canvas panel, using traditional art methods including a tracing paper guide featured earlier. This work shows in retrospect my ideal of the 'spirit of space art', showing something of a summary of what is known of the subject while conveying what it is about the subject's known or theorized attributes that I find visually compelling. The new digital painting updates that visual impression with one based on our current knowledge. The new digital work was also done at the same size as this one was for possible printing purposes, 16 by 20 inches at 300 DPI. #Sciart In 1976 the pair of Viking spacecraft orbiter/lander combos were sent to Mars. Before Viking, Mars was still largely a place of our imaginations, an unknown that we wrote stories about then studied from Earth and space. The dominance of cratered terrain on the portions of Mars photographed by Mariners 4, 6 and 7 gave an initially Lunar impression of that world, which changed during the development of the Viking lander after Mariner 9 revealed the planet. The first landing was aimed for a flat flood plain, judged as a decent place to search for life. The landing date was July 4, however the intended site appeared too rough for comfort in the high quality orbiter photos, causing the landing to wisely be delayed. A smoother site further 'down stream' was found using photo and radar data. The landing proceeded quite nominally, a magnificent achievement that apparently accounted for what it had to and made it the first time. The first photo from the surface was a look at a footpad to see how solid the 'footing' appeared. Then a moderate resolution panorama of the surroundings was obtained, first showing the rocks, hills, dunes-and bright sky. The latter was my biggest surprise as a space artist. The adjustments to the color balance of the later photos were gradually narrowed down. The ubiquitous dust is planet wide, suspended in the air and coloring the sky a luminous tan ochre. This digital painting shows the view a meter above the surface seeing the lander briefly kicking up dust less than a second before touchdown with the footpads 1.4 meters high. The landing rockets were designed with many nozzles to disturb the surroundings as little as possible. The nearby disturbed dust resembles in color that of the sky. We are 6 meters away from the center of the lander. field of view is 60 degrees, we are looking Northwest with the Sun and Earth off the frame to the left. This is a digital art update of the pre mission oil painting I did for NASA HQ 1n 1975. #Sciart
Tom Phillips@tom-phillips.com · 3d ago

It is quite strange how little the actual public inquiry we're having into the UK's Covid response - the one that's publishing in-depth reports every few months - factors into the "you're not even allowed to ask these questions" complaints. We are, in fact, officially asking these questions!

Zack Whittaker@zackwhittaker.com · 5d ago

Keen to see if any of the companies that were hacked by OpenAI or Anthropic will sue them. Someone has to take responsibility for this, and the blame is almost entirely on the leaders of these AI companies. Alternatively, hacking is just legal now until a court says otherwise? What a fucking mess.

Daragh Ó Briain@daraghobrien.bsky.social · 6d ago

Sorry: they only did reviews to see if their software had unlawfully and without authorisation accessed networks of third parties? This wasn’t a defined control *during* their “testing”? This is extreme negligence at least. cyberscoop.com/anthropic-cl...

Jacob Aron@jjaron.bsky.social · 5d ago

I do genuinely appreciate Badenoch attempting to make the distinction between "net zero", a fact of atmospheric physics, and "net zero by 2050", a political programme. Unfortunately she has done this quite poorly, but it's something that we should be encouraging critics of "net zero" to spell out

James Murray@james-bg.bsky.social · 6d ago

This is even worse than Badenoch’s letter yesterday, as she has space to explain what she would do instead of net zero targets to tackle the climate change she accepts is real but then doesn’t. She accuses Labour of not having a plan, then insists we should trust her unannounced plan will be better.

New Scientist@newscientist.com · 5d ago

To predict when we’ll have a truly useful quantum computer, we need to know precisely what one is. Columnist Karmela Padavic-Callaghan finds that’s harder to pin down than you might think

Why full-fledged quantum computers might always be five years away | New Scientist

To predict when we’ll have a truly useful quantum computer, we need to know precisely what one is. Columnist Karmela Padavic-Callaghan finds that’s harder to pin down than you might think

newscientist.com

Simon Evans@drsimevans.carbonbrief.org · 6d ago

🧵UK media has been duped about net-zero (again) The playbook is identical: 1️⃣ Find a big scary number 2️⃣ Speed-dial non-specialist hacks 3️⃣ Do say “serious economic damage” 4️⃣ Don’t say “we pretended gas is free” 5️⃣ That's it! Many papers still fell for it 🤷

Bild