In April Anthorpic introduced a new AI trick: multiple Claude agents combine for a "Research" feature that can "search across both your internal work context and the web" (as well as Google Workspace "and any integrations...")

But a recent Anthropic blog post notes this feature "involves an agent that plans a research process based on user queries, and then uses tools to create parallel agents that search for information simultaneously," which brings challenges "in agent coordination, evaluation, and reliability.... The model must operate autonomously for many turns, making decisions about which directions to pursue based on intermediate findings." Multi-agent systems work mainly because they help spend enough tokens to solve the problem.... This finding validates our architecture that distributes work across agents with separate context windows to add more capacity for parallel reasoning. The latest Claude models act as large efficiency multipliers on token use, as upgrading to Claude Sonnet 4 is a larger performance gain than doubling the token budget on Claude Sonnet 3.7. Multi-agent architectures effectively scale token usage for tasks that exceed the limits of single agents.

There is a downside: in practice, these architectures burn through tokens fast. In our data, agents typically use about 4Ã-- more tokens than chat interactions, and multi-agent systems use about 15Ã-- more tokens than chats. For economic viability, multi-agent systems require tasks where the value of the task is high enough to pay for the increased performance. Further, some domains that require all agents to share the same context or involve many dependencies between agents are not a good fit for multi-agent systems today.

For instance, most coding tasks involve fewer truly parallelizable tasks than research, and LLM agents are not yet great at coordinating and delegating to other agents in real time. We've found that multi-agent systems excel at valuable tasks that involve heavy parallelization, information that exceeds single context windows, and interfacing with numerous complex tools.
Thanks to Slashdot reader ZipNada for sharing the news.

Long-time Slashdot reader lee1 shares a new article from Linux magazine: The danger and irritations of the modern web have unleashed a movement dedicated to creating a safer and simpler alternative. The old Gopher network and the new Gemini protocol have emerged as building blocks for this new "small Internet."

Anyone who has used the World Wide Web (WWW) lately knows that something bad is happening to it. It does not resemble the WWW of the early years, with enthusiastic amateurs freely sharing ideas and information. These things still exist, and the web is still an indispensable medium connecting the world. But the web experience is now encumbered with advertising, invasions of privacy in the form of pervasive tracking, enormous file sizes, CPU straining JavaScript, the danger of exploits, and door slams asking you to subscribe to a newsletter before viewing a site.

This unpleasant environment has led to a backlash. There are now some communities of developers and computer users who still desire a connected information system, but who seek a refuge from the noise, danger, and increasingly resource-hungry WWW. They feel that web technology does too much, and that since it makes various forms of abuse too easy, no lasting reform is possible.

The solution is to use or create a separate protocol that is simply not capable of supporting the technologies that enable advertising networks, user fingerprinting, or the myriad of other things that exploit users rather than helping them. This small movement has approached the problem from two directions that in practice are often merged: the revival of the Gopher protocol and the creation of a new protocol called Gemini.
Gemini would support its own lightweight hypertext format, and would co-exist with Gopher and HTTP as an alternative client-server protocol with built-in privacy-assuring features like mandatory Transport Layer Security and a "Trust On First Use" public-key security model. ("Connections are closed at the end of a single transaction and cannot be reused," notes the Project Gemini home page.) "You may think of Gemini as 'the web, stripped right back to its essence,'" explains its FAQ, "or as 'Gopher, souped up and modernised just a little', depending upon your perspective..."

"Gemini is also intended to be very privacy conscious, to be difficult to extend in the future (so that it will *stay* simple and privacy conscious), and to be compatible with a 'do it yourself' computing ethos."

sciencehabit shares an excerpt from Science Magazine: Talk about getting something for nothing. Physicists predict that just by shooting charged particles through an electromagnetic field, it should be possible to generate light from the empty vacuum. In principle, the effect could provide a new way to test the fundamental theory of electricity and magnetism, known as quantum electrodynamics, the most precise theory in all of science. In practice, spotting the effect would require lasers and particle accelerators far more powerful than any that exist now. Physicists have long known that energetic charged particles can radiate light when they zip through a transparent medium such as water or a gas. In the medium, light travels slower than it does in empty space, allowing a particle such as an electron or proton to potentially fly faster than light. When that happens, the particle generates an electromagnetic shockwave, just as a supersonic jet creates a shockwave in air. But whereas the jet's shockwave creates a sonic boom, the electromagnetic shockwave creates light called Cherenkov radiation. That effect causes the water in the cores of nuclear reactors to glow blue, and it's been used to make particle detectors.

However, it should be possible to ditch the material and produce Cherenkov light straight from the vacuum, predict Dino Jaroszynski, a physicist at the University of Strathclyde in Glasgow, U.K., and colleagues. The trick is to shoot the particles through an extremely intense electromagnetic field instead. According to quantum theory, the vacuum roils with particle-antiparticle pairs flitting in and out of existence too quickly to observe directly. The application of a strong electromagnetic field can polarize those pairs, however, pushing positive and negative particles in opposite directions. Passing photons then interact with the not-quite-there pairs so that the polarized vacuum acts a bit like a transparent medium in which light travels slightly slower than in an ordinary vacuum, Jaroszynski and colleagues calculate. Putting two and two together, an energetic charged particle passing through a sufficiently strong electromagnetic field should produce Cherenkov radiation, the team reports in a paper in press at Physical Review Letters. Others had suggested vacuum Cherenkov radiation should exist in certain situations, but the new work takes a more fundamental and all-encompassing approach, says Adam Noble, a physicist at Strathclyde.

With the holidays coming up, Bennett Haselton has updated his geek-oriented gift guide for 2014. He says: Some of my favorite gifts to give are still the ones that were listed in several different previously written posts, while a few new cool gift ideas emerged in 2014. Here are all my current best recommendations, listed in one place. Read on for the list, or to share any suggestions of your own.

Bennett Haselton writes My LG Optimus F3Q was the lowest-end phone in the T-Mobile store, but a cheap phone is supposed to suck in specific ways that make you want to upgrade to a better model. This one is plagued with software bugs that have nothing to do with the cheap hardware, and thus lower one's confidence in the whole product line. Similar to the suckiness of the Stratosphere and Stratosphere 2 that I was subjected to before this one, the phone's shortcomings actually raise more interesting questions — about why the free-market system rewards companies for pulling off miracles at the hardware level, but not for fixing software bugs that should be easy to catch. Read below to see what Bennett has to say.

Hugh Pickens DOT Com writes "David Talbot writes at MIT Technology review that engineers on the Internet Engineering Task Force (IETF), an informal organization of engineers that changes Internet code and operates by rough consensus, have asked the architects of Tor to consider turning the technology into an Internet standard. If widely adopted, such a standard would make it easy to include the technology in consumer and business products ranging from routers to apps and would allow far more people to browse the Web without being identified by anyone who might be spying on Internet traffic. The IETF is already working to make encryption standard in all web traffic. Stephen Farrell believes that forging Tor into a standard that interoperates with other parts of the Internet could be better than leaving Tor as a separate tool that requires people to take special action to implement. 'I think there are benefits that might flow in both directions,' says Farrell. 'I think other IETF participants could learn useful things about protocol design from the Tor people, who've faced interesting challenges that aren't often seen in practice. And the Tor people might well get interest and involvement from IETF folks who've got a lot of experience with large-scale systems.' Andrew Lewman, executive director of Tor, says the group is considering it. 'We're basically at the stage of 'Do we even want to go on a date together?' It's not clear we are going to do it, but it's worth exploring to see what is involved. It adds legitimacy, it adds validation of all the research we've done.'"

Frequent contributor Bennett Haselton writes this week with his favorite novelty science gift items for 2012. Levitation engines, puzzles, optical illusions brought to life, and all of the tips and tricks he's found for getting the products to work correctly. Decorative, whimsical, and not too expensive — except for the items that have earned it by being pretty amazing. Read on for the details, and be sure to mention other good possibilities (Just 14 shopping days left until Christmas) in the comments below.

Frequent Slashdot contributor Bennett Haselton writes "An estimated 200,000 Hotmail users currently have their auto-reply set to a message spamming an advertisement for Chinese scam websites, which sell "discounted" electronics. Presumably the spammers compromised a large number of Hotmail accounts to pull this off, but wouldn't it be pretty easy for Hotmail to query for which users have that set as their auto-reply, and turn the auto-reply off for them?" Read below for Bennett's thoughts.

gregrolan writes "The Trousers of Reality — Volume 1, Working Life is indeed a book about finding balance and satisfaction in life work and play. The author's thesis can be applied to almost any discipline, but it is from his background as an IT consultant that most of his professional examples are drawn. He considers success in this field pretty broadly and addresses the technical, management, political, personal, and social aspects of the IT profession." Read on for the rest of Greg's review.

Beginning with the original 1993 SNES title, Nintendo has toyed with the space combat series Star Fox in a number of ways. Star Fox Adventures added a doodad-hunting adventure format, while last year's Star Fox Assault included ground-based battles that managed to actually detract from the game's appeal. Now on the DS, the Star Fox series has experienced yet another format shift: Star Fox Command. Command offers some simple strategy elements, an innovative control scheme, and the tried and true dogfighting gameplay the series is known for. It also dwells on one of the series' weaknesses, plot, to the detriment of the game. Despite that and a few other issues, Star Fox Command will provide some quality space combat for anyone looking to kill some time, and a few aliens. Read on for my impressions of Nintendo's latest core series release.

Max Spevack writes: "Hi everyone. I'm looking forward to answering all of the questions, but before I start diving into that, I guess it would be useful to give a little bit of perspective about me and my role within Fedora and Red Hat, because it will offer some context around the things I have to say."

The unique elements of Nintendo's Dual Screen handheld have led to titles with very different gameplay. Warioware Touched and Kirby Canvas Curse typify the ways that Nintendo wants game designers to begin thinking about using their hardware. There are still new directions that Nintendo wants to push gaming, though, and they're not content to sit back and let others find the path. Non-game games, then, are what seems to be in store for the DS. Titles like Electroplankton, the music-making system, and the tamagotchi-esque Nintendogs are intended to bring non-gamers into the fold with interactive software that they can enjoy but aren't necessarily games. In the spirit of the non-game games, then, read on for my non-review of Nintendogs.

sympleko (Matthew Leingang) writes "In Neal Stephenson's manifesto In the Beginning was the Command Line , he writes about his favorite command-line utility: wc. As simple as can be, wc counts characters, words, and lines in a file. There's no GUI analogue, perhaps because anybody tempted to make one would add too many "features" that cluttered its ease of use. Think: do you know how to count the words in a Word file? BBEdit is a little easier, if you know the button to click." Read on for Leingang's review of Learning Unix for Mac OS X Panther, which seeks to reconcile the conception of user friendliness in OS X's Aqua with the sometimes-denigrated command line.

Slashdot readers both contributed and helped moderate questions for the New Voters Project Presidential Youth Debate. You can read the answers below, but if you'd like to see an expanded introduction, thumbnails of the candidates, and different formatting, go to the Youth Debate page. And that's not all: We're supposed to get candidates' rebuttals on or about October 17, so don't touch that dial!

Joel Natt writes "Planet Broadband is not a Star Trek episode or another Sci-Fi title. It is the title of a new book by Rouzbeh Yassini which answers the question of where the term 'broadband' originated and why is it used when discussing cable internet services." Read on for the rest of Natt's review of Planet Broadband.

Chris Cowell-Shah writes "If David Foster Wallace can't explain infinity to us, nobody can. At least, that's what I told myself while anxiously waiting for his Everything and More: A Compact History of Infinity. The book promised to be an intellectual history of the mathematical concept of infinity, with heavy doses of history, math, and philosophy. And while it proves heavy going at times, I'm pleased to say that it delivers admirably on this promise." Read on for Cowell-Shah's lengthy review of Everything and More.

Some excellent questions got asked. And these answers, from Larry Rosen, an attorney who works heavily on open source licensing matters, ought to give you a bit of insight into what the Microsoft "final judgement" means in the context of open source development and the software marketplace in general.

Major news Web sites routinely rewrite stories after they are published, sometimes so heavily that they only bear a glancing resemblance to what was posted earlier. This CNN/Money article about the penalty phase of the Microsoft trial is a prime example. What you see at the other end of the link is quite different from the story that first appeared at that URL. Even the headline and byline have changed. But CNN/Money managing editor Allen Wastler says there is nothing wrong with this practice, even though there is no indication on the site that the article was heavily modified after it first appeared.

Reader Torulf contributed the below review of Jef Raskin's The Humane Interface .Though the book does not spend much time on Open Source software, it emphasizes ideas that every programmer probably ought to bear in mind -- at least if they wants hisprograms to have users. (And yes, he takes explicit exception to some UNIX truisms.)

Okay, the answers to your questions for U.S. Representative Rick Boucher are in. No, his staff didn't write them. Everything you see here is straight from the Congressman himself. This is a nice bit of insight into legislative thinking about the Internet, and gives a little info on how you can help change laws you don't like, too.