Penny Wong, Foreign Minister: Good morning, I'll just make a few brief comments. First, in relation to what is occurring in the Middle East, Australia has been clear along with the international community that Iran cannot be allowed to develop a nuclear weapon. We support action to prevent that from occurring. And this is what this is. What I would also say is we know that the international nuclear watchdog, the International Atomic Energy Agency, has made clear that Iran has enough heavily enriched uranium for several nuclear bombs. That has been made clear. It has also said that Iran is not in compliance with its obligations under the non-proliferation arrangements that it has signed up to. What happens now matters. We do not want to see escalation. We do call for diplomacy, de-escalation and dialogue because the world does not want to see full-scale war in the Middle East. And that call from Australia is consistent with the same call from the United Kingdom, members of the European Union and beyond. A brief consular update – we currently have 2,900 Australians and their families in Iran who have registered with us and are seeking to leave. We have around 1,300 Australians and their families who are still registered with us in Israel and are seeking to leave. The situation is very difficult. Airspace remains closed, certainly in Iran it is a very difficult situation. You have heard me previously urging Australians and their families to leave if they believe they can do so safely. We've deployed Australian officials to the Azerbaijani border, so if Australians do get there, we can provide them with support. In relation to Israel, there are reports that the airspace may open for a limited period. Obviously this is very fluid, but we are seeking to make arrangements to utilise that window if we are able. And we have advised Australians on the ground of that fact. Again, I emphasise we are seeking to utilise this opportunity but the situation on the ground is uncertain and fluid and risky. I'm happy to take questions.

Journalist: Did Australia receive any advance notice that the US would be taking this action? And what role could Australia play in the de-escalation process aside from public advocacy, as you are today?

Foreign Minister: Well, the US has made clear this was unilateral action taken by the United States. What I would say is we're not a central player. What we can do is to add our voice to the calls from the international community who, overwhelmingly, are calling for de-escalation and diplomacy.

Journalist: There's a chorus of international lawyers since this happened that have said that this is a breach of international law, it's not in keeping with the UN Charter. How do you justify these actions?

Foreign Minister: I would make the point that the facilities which have been targeted are only for the purpose of Iran's nuclear program. I would make that point and I think that does matter. We have long been clear, along with others in the international community that it is a threat to global peace and security for Iran to obtain a nuclear weapon.

Journalist: Why has it taken almost 24 hours for the government to provide support to the US?

Foreign Minister: I think if you look at our statement released by the Government yesterday, we made clear what we have always said, which is the international community including Australia has long believed that Iran cannot obtain a nuclear weapon. I think that was very clear.

Journalist: What do you say to Iranian community in Australia about Australia's support I suppose?

Foreign Minister: Look, I understand this is a very difficult time for the Australian-Iranian community. The conflict itself, their concern about friends and family in Iran, including Australians and their families who may be in Iran. It is a very difficult time and we are thinking of you. What I can say, certainly in relation to Australians and their families, is that we stand ready to assist them when we are able to.

Journalist: One question about the assisted departures – so those numbers across Iran and Israel is about 4,200 now?

Foreign Minister: That's right.

Journalist: Assuming that the airspace does open and you acknowledge that the situation is fluid, how long could it potentially take?

Foreign Minister: I'm not going to speculate on that because, you know, there are so many, there are so many variables in play. But again, what I'd say is we will seek to assist when we can and wherever we can. The situation is unstable and the situation is fluid. Thank you very much.

Apologies for the non-appearance of Wednesday’s post, and Friday before that – life intervened in the form of a bunch of conference travel. While at a finance conference this week, I got to thinking about the differences between private and public sessions, which in turn reminded me of an invention which I came up with at another conference last year, which I’m sure will make my fortune as soon as I can vibe-code it.

This invention, which is surely worth billions, takes the form of an add-in for Slack, Teams or similar corporate chat and communication systems. It’s called “The I Never Said That Button”.

As the name suggests, you click the button and a text field pops up for you to enter a short description of a concept, view or implied promise which you would currently prefer to have have never said, but which you suspect that the Slack or email archive may contain text which kind of looks like you did say it. Using LLM technology, the add-in scours the complete record, deletes anything that looks like the thing you never said, then replaces it with automatically generated innocuous and noncommittal text.

Obviously, there will be disclaimers needed to warn against its use in regulated financial institutions or government departments subject to Freedom of Information requests. And equally obviously, this isn’t a mass market product; I was thinking of starting the pricing at $10,000 and restricting the licences so that in any organisation, you can be reasonably sure that only the boss has one. (“Only the boss has one” is a key part of the value offering of course, and I’m thinking of trademarking that phrase as a subsidiary marketing slogan).

I joke, I joke of course.

But the policy of this ‘stack is always to take jokes seriously and metaphors literally, because you will often be surprised at what you learn. And in this case, it is interesting to consider why an “I Never Said That Button” might be a joke worth telling.

Denying that you said something, when you did, is a sin as old as management itself; it is a core function of being a boss. Lots of leadership styles rely on projecting a level of certainty and commitment which really isn’t consistent with the tendency of the world to change.

I don’t think anyone necessarily paid attention to the possible consequence of making this kind of normal management behaviour so much more difficult, when we moved to a world of largely text-based communication.

And it’s also affected the communication itself. Anyone who’s been active on social media for a while is likely to have on occasion caught themselves in the act of rephrasing something defensively. There are plenty of times when it’s appropriate to write on the basis that your words will be given the worst possible interpretation by your worst enemy at the worst possible moment. There just didn’t used to be so many use of them.

The strategic use of communication is something I’m going to be writing quite a bit more about over the next few weeks, because it’s a subset of strategic behaviour more generally, and I think that if we look hard at this kind of subject, we might find the source of quite a lot of our problems of governance and state capacity.

I wrote about “Pre-emptive risk aversion and state capacity” back in April, and I’m beginning to think the concept might generalise. The problem with the regulatory state and abundance might not be the regulations themselves, but the way in which the regulatory system facilitates strategic behaviour. (A major factor in the Sheephouse Woods Bat Structure, our favourite example, was the utterly cynical use of Tree Protection Orders by Buckinghamshire Council).

And since I’m on record in that context as saying that a huge amount of the problem of state capacity is caused by the absence of informal communication channels which can be used to discuss problems and solutions before commitments and decisions are made, I think the fact that we’ve structurally directed public regulatory bodies to only use archived text – the very worst possible medium for informal or preparatory discussions – could be very relevant.

Maybe Stafford Beer was on to something when he argued that the key tool of management cybernetics was a room full of cigars, comfortable armchairs and whisky.

Click here to go see the bonus panel!

Hovertext:
I feel like I could at least respect if political leaders would knock down legal barriers with a folding chair, instead of just ignoring them.

Today's News:

FOR IMMEDIATE RELEASE
Saint John, NB – June 2025

At Canada’s leading technical nuclear conference this week, Moltex Energy proudly unveiled major breakthroughs that could redefine how nuclear waste is handled — turning used fuel into a valuable resource while sharply reducing the volume of long-term waste.

---

Key Announcements from Moltex Energy Canada at the Canadian Nuclear Society Conference

Recycling Used Fuel

CEO Rory O’Sullivan delivered a plenary presentation on the exciting progress of Moltex’s Waste-to-Stable-Salt (WATSS) process, highlighting the growing economic and environmental value of used nuclear fuel.

Based on current stockpiles in North America alone, there is:

• $80 billion in fuel value from transuranic elements
• $60 billion in residual uranium
• $30 billion in rare earth elements

“This is a game-changer for the sustainability of nuclear energy,” said O’Sullivan.
“We’re showing that nuclear power can be not only carbon-free but resource-smart — turning yesterday’s waste into tomorrow’s clean energy. It is no longer acceptable to bury such a valuable resource.”

---

Validated by First-of-Its-Kind Testing

Lead chemist Jose P. Zuniga shared groundbreaking results from real-world testing on used CANDU reactor fuel, conducted in collaboration with Canadian Nuclear Laboratories:

• Successfully extracted over 77% of the transuranics from real used fuel — a major scientific and commercial milestone
• New testing shows over 95% extraction, to be presented at the
  ANS Advances in Nuclear Fuel Management Conference,
  July 20–23, 2025, Clearwater, Florida

---

Reducing Long-Term Nuclear Waste

Chief Technology Officer Olivier Grégoire described how the WATSS process can reduce the footprint of long-term nuclear waste repositories by as much as 80%. The process enables:

• Advanced borehole disposal technologies
• Reduced surface dose levels
• Significantly lower waste storage requirements

These innovations offer more compact, flexible, and cost-effective disposal solutions.

---

Engaging with Global Experts

Moltex also participated in the Plenary Session on Advances in Reactor Technologies, where the team discussed how next-generation fuel recycling can integrate with modern reactor designs.

---

Acknowledging the Nuclear Community

Moltex Energy Canada expressed its sincere appreciation to the Canadian Nuclear Society for providing a global platform to share these critical advancements in nuclear science and technology.

---

About Moltex Energy Canada

Moltex Energy Canada, part of the Moltex Energy group, is developing advanced nuclear technologies designed to safely recycle existing spent fuel into new energy. Headquartered in Saint John, New Brunswick, Moltex is helping Canada lead the world in clean, innovative nuclear solutions.

---

Media Contact
Betty Draper
Office Manager
Moltex Energy Canada
bettydraper@moltexenergy.com
506-721-9551

The post Moltex Energy Unveils Breakthrough in Nuclear Waste Recycling at Canadian Nuclear Society Conference appeared first on Moltex Energy.

MIT graduate student Alex Kachkine once spent nine months meticulously restoring a damaged baroque Italian painting, which left him plenty of time to wonder if technology could speed things up. Last week, MIT News announced his solution: a technique that uses AI-generated polymer films to physically restore damaged paintings in hours rather than months. The research appears in Nature.

Kachkine's method works by printing a transparent "mask" containing thousands of precisely color-matched regions that conservators can apply directly to an original artwork. Unlike traditional restoration, which permanently alters the painting, these masks can reportedly be removed whenever needed. So it's a reversible process that does not permanently change a painting.

"Because there's a digital record of what mask was used, in 100 years, the next time someone is working with this, they'll have an extremely clear understanding of what was done to the painting," Kachkine told MIT News. "And that's never really been possible in conservation before."

Figure 1 from the paper. Credit: MIT

Nature reports that up to 70 percent of institutional art collections remain hidden from public view due to damage—a large amount of cultural heritage sitting unseen in storage. Traditional restoration methods, where conservators painstakingly fill damaged areas one at a time while mixing exact color matches for each region, can take weeks to decades for a single painting. It's skilled work that requires both artistic talent and deep technical knowledge, but there simply aren't enough conservators to tackle the backlog.

The mechanical engineering student conceived the idea during a 2021 cross-country drive to MIT, when gallery visits revealed how much art remains hidden due to damage and restoration backlogs. As someone who restores paintings as a hobby, he understood both the problem and the potential for a technological solution.

To demonstrate his method, Kachkine chose a challenging test case: a 15th-century oil painting requiring repairs in 5,612 separate regions. An AI model identified damage patterns and generated 57,314 different colors to match the original work. The entire restoration process reportedly took 3.5 hours—about 66 times faster than traditional hand-painting methods.

Alex Kachkine, who developed the AI-printed film technique. Credit: MIT

Notably, Kachkine avoided using generative AI models like Stable Diffusion or the "full-area application" of generative adversarial networks (GANs) for the digital restoration step. According to the Nature paper, these models cause "spatial distortion" that would prevent proper alignment between the restored image and the damaged original.

Instead, Kachkine utilized computer vision techniques found in prior art conservation research: "cross-applied colouration" for simple damages like thin cracks, and "local partial convolution" for reconstructing low-complexity patterns. For areas of high visual complexity, such as faces, Kachkine relied on traditional conservator methods, transposing features from other works by the same artist.

From pixels to polymers

Kachkine's process begins conventionally enough, with traditional cleaning to remove any previous restoration attempts. After scanning the cleaned painting, the aforementioned algorithms analyze the image and create a virtual restoration that "predicts" what the damaged areas should look like based on the surrounding paint and the artist's style. This part isn't particularly new—museums have been creating digital restorations for years. The innovative part is what happens next.

Custom software (shared by Kachkine online) maps every region needing repair and determines the exact colors required for each spot. His software then translates that information into a two-layer polymer mask printed on thin films—one layer provides color, while a white backing layer ensures the full color spectrum reproduces accurately on the painting's surface. The two layers must align precisely to reproduce colors accurately.

High-fidelity inkjet printers produce the mask layers, which Kachkine aligns by hand and adheres to the painting using conservation-grade varnish spray. Importantly, the polymer materials dissolve in standard conservation solutions, allowing future removal of the mask without damaging the original work. Museums can also store digital files documenting every change made during restoration, creating a paper trail for future conservators.

Kachkine says that the technology doesn't replace human judgment—conservators must still guide ethical decisions about how much intervention is appropriate and whether digital predictions accurately capture the artist's original intent. "It will take a lot of deliberation about the ethical challenges involved at every stage in this process to see how can this be applied in a way that's most consistent with conservation principles," he told MIT News.

For now, the method works best with paintings that include numerous small areas of damage rather than large missing sections. In a world where AI models increasingly seem to blur the line between human- and machine-created media, it's refreshing to see a clear application of computer vision tools used as an augmentation of human skill and not as a wholesale replacement for the judgment of skilled conservators.

Read full article

Comments

They poison their own context. Maybe you can call it context rot, where as context grows and especially if it grows with lots of distractions and dead ends, the output quality falls off rapidly. Even with good context the rot will start to become apparent around 100k tokens (with Gemini 2.5).

They really need to figure out a way to delete or "forget" prior context, so the user or even the model can go back and prune poisonous tokens.

Right now I work around it by regularly making summaries of instances, and then spinning up a new instance with fresh context and feed in the summary of the previous instance.

— Workaccount2 on Hacker News, coining "context rot"

Tags: long-context, llms, ai, generative-ai