Keith Smith

Early universe’s ‘little red dots’ may be black hole stars Puzzling objects spotted by NASA’s JWST telescope may be entirely new class of celestial entity

Early universe’s ‘little red dots’ may be black hole stars | Science | AAAS www.science.org/content/arti... 🔭🧪 @science.org

Professor Michele Dougherty.

The Royal Astronomical Society is pleased to congratulate Professor Michele Dougherty on her appointment as the first female Astronomer Royal. 🔭💫🪐

Read more at: ras.ac.uk/news-and-pre...

⤵️

The prompts in the news story were not trying to uncover illicit LLM use. They were trying to *exploit* it to give themselves an unfair advantage.

I'm not objecting to all injections, I'm objecting to those designed to benefit the authors while disadvantaging everyone else.

That's not what I said (or assumed) at all.

LLMs should not be used to review a paper; doing so is unethical for a large number of reasons. That does not justify further unethical behaviour - like the 'give a positive review' prompts reported in the news article that started this thread.

Our instructions to referees explicitly ban their use, see point 5 www.science.org/content/page...
Determining whether referees have followed those instructions is trickier, as with the other ethical guidelines.

We do have software that can identify hidden prompts in manuscripts.

bsky.app/profile/drke...

It's not "potentially questionable", it's highly unethical. Two wrongs do not make a right, as every child is taught. Fixing a positive review is not a way of catching LLMs, it's a way of benefiting from their misuse.

A catch word or phrase would not benefit the authors, so is far less concerning.

No, if that prompt works it also has an effect on every other author who didn't try to game the system that way, and on the conference/journal that now gets even less useful reports.

The hypothetical LLM-using referee won't be hurt at all, but the overall peer review system is damaged even further.

I agree that referees should not be using LLMs in the first place. But if one did, then including this prompt a) makes the LLM problem much worse, and b) does so in a way that unfairly benefits the authors. That's the unethical bit.

Inserting 'artichoke' would not be a benefit, so less concerning.

Because these prompts attempt to undermine the peer review process and avoid assessment of the scientific contents. Trying to covertly force a positive report is unethical, regardless of whether the attempt works.

The ICML likens it to offering a bribe for a positive review.
icml.cc/Conferences/...

🤦‍♂️

The solution to unethical behaviour (using AI to review papers) is not more unethical behaviour (inserting hidden AI prompts).

'Positive review only': Researchers hide AI prompts in papers Instructions in preprints from 14 universities highlight controversy on AI in peer review

A number of authors have been caught including hidden prompts directed at any AI tools used to review their papers. 🧪☄️

🤦‍♂️

The solution to unethical behaviour (using AI to review papers) is not more unethical behaviour (inserting hidden AI prompts).

Unusual unit of the day: analytic sensitivity measured in yoctograms per gram.

That's equivalent to 1 part in 10^24, roughly 1.7 atoms or molecules per mole (for a fixed atomic or molecular weight). 🧪⚛️ #chemsky

An offer of €400m is a nice problem to have. I hope they take it.

China is quietly preparing to build a gigantic telescope Astronomers are puzzled by the silence over an instrument that could briefly reign as world’s biggest

LOT hopes to reach first light before E-ELT does, so *might* hold the record for a few months. Hence the question mark. 🔭

Bob Kirshner is talking up that possibility to drum up funding for the TMT. That’s understandable, but over-emphasises competition.
www.science.org/content/arti...

Editor's choice: Roche et al. simulate how giant impacts can remove a planet's atmosphere. They simulate hundreds of impact configurations, from which they derive a scaling law for what fraction of the atmosphere is lost in each impact. ☄️ #planetsci #exoplanet
www.science.org/doi/10.1126/...

That image was the cover of @science.org on 16 March 2016. That issue contained five papers by the New Horizons team reporting results from the flyby of Pluto.
www.science.org/toc/science/...

I’m not an expert in speckle imaging, but “1.5σ detection” sounds less like a detection and more like a hopeful rumor.🔭

The Probable Direct-Imaging Detection of the Stellar Companion to Betelgeuse Betelgeuse -- the closest M-supergiant to the Sun -- has recently been predicted to host a lower-mass stellar companion that orbits the primary with a period of $\sim 6$ years. The putative stellar co...

It’s in the abstract of arxiv.org/abs/2507.15749

😮 Just about everyone who has ever submitted an observing proposal understands the concept of non-sidereal tracking, if only so they can be confident in skipping that section of the form...

The US Senate is not willing to go along with Trump’s proposed cancellation of dozens of NASA missions. 🧪🔭 #planetsci

Feeds - The Astrosky Ecosystem We're building social media tools for the astronomy & space science communities. From feeds to hosting, we're billionaire-proofing scientific discussion for good.

- Is there a concise up-to-date list of all the symbols & hashtags used to post to each feed? astrosky.eco/feeds requires opening each one individually to see them.

- The new pinned posts have pretty images but take up the entire screen (mobile) or browser window (desktop), which seems excessive.

(The exception is Japan, and presumably this statistic is pairing ESA's launch capability with UK & France nuclear weapons)

A sobering statistic that I heard in @bleddb.bsky.social's talk at #NAM2025 last week:
There are 92 countries & companies with active spacecraft, but only ten with launch capabilities; nine of those have nuclear weapons. Therefore access to space is reliant on the military-industrial complex. 🧪🚀🛰️🔭

A high-resolution color image of Mars with the small, dark, potato shaped satellite Phobos, seen in sharp contrast as it passes over the massive volcano Olympus Mons. Phobos appears irregular and shadowed against the softly rust colored Martian surface. In the distance, two other volcanic features Ascraeus Mons (upper right) and the flatter Alba Mons (upper left) are visible. Wisps of light clouds and the blackness of space frame the planet’s curved horizon.

1/n 🧵
Just released: NEW images of Phobos over Mars by @esa.int Mars Express 🔭🧪

This view shows Phobos above Olympus Mons!

Full-size (130MP) image & details: flic.kr/p/2rggHKy
Credit: ESA/DLR/FU Berlin/Andrea Luck CC BY

Captured on May 14, 2025 | Image ID: HQ967
Raw data from: psa.esa.int

Astronomers race to study interstellar interloper Phalanx of instruments scrutinize 3I/ATLAS for clues about planet formation elsewhere in the galaxy

Discovered last week, a third interstellar visitor is whizzing through the Solar System

Adam Mann has been speaking to researchers about what they hope to learn from it - and how quickly they need to react. ☄️ #planetsci

www.science.org/content/arti...

Opps! I’ll fix that, thanks for alerting me

On 1 July, astronomers discovered 3I/ATLAS, just the third interstellar object seen passing through our Solar System.

The Very Large Telescope in Chile recently obtained this 13-minute time lapse of it moving against a background of distant stars: scim.ag/44NLult

U.S. abandons hunt for signal of cosmic inflation Now-canceled CMB-S4 project would have searched the afterglow of the Big Bang for signs of cosmic exponential growth spurt

Breaking news: The U.S. government has canceled a proposed $900 million project to study in unprecedented detail the afterglow of the Big Bang, the cosmic microwave background radiation. scim.ag/4eIo0m4