Microsoft has introduced BitNet b1.58 2B4T, the largest-scale 1-bit AI model to date with 2 billion parameters and the ability to run efficiently on CPUs. It's openly available under an MIT license. TechCrunch reports: The Microsoft researchers say that BitNet b1.58 2B4T is the first bitnet with 2 billion parameters, "parameters" being largely synonymous with "weights." Trained on a dataset of 4 trillion tokens -- equivalent to about 33 million books, by one estimate -- BitNet b1.58 2B4T outperforms traditional models of similar sizes, the researchers claim.

BitNet b1.58 2B4T doesn't sweep the floor with rival 2 billion-parameter models, to be clear, but it seemingly holds its own. According to the researchers' testing, the model surpasses Meta's Llama 3.2 1B, Google's Gemma 3 1B, and Alibaba's Qwen 2.5 1.5B on benchmarks including GSM8K (a collection of grade-school-level math problems) and PIQA (which tests physical commonsense reasoning skills). Perhaps more impressively, BitNet b1.58 2B4T is speedier than other models of its size -- in some cases, twice the speed -- while using a fraction of the memory.

There is a catch, however. Achieving that performance requires using Microsoft's custom framework, bitnet.cpp, which only works with certain hardware at the moment. Absent from the list of supported chips are GPUs, which dominate the AI infrastructure landscape.

An anonymous reader writes: On June 29, Microsoft announced that Windows 10 was running on 350 million devices -- 50 million more devices than the previous milestone announced by Microsoft on May 5. While the company is expected to update the number of devices running the latest OS when it releases the Windows 10 Anniversary Update on August 2nd, NetMarketShare has decided to conduct some research on its own. According to its report, Windows 10 currently runs on a 21.13% desktop OS share. Meanwhile, Windows 7 continues to dominate the market with a 47.01% share, with Windows 8 and Windows 8.1 representing less than 10% of the PC market, and Windows XP representing 10.34%. While the market share of Windows 10 is all but certain to rise, it likely won't rise as fast as it did between May and June or June and July for example, as Windows 10 is no longer offered as a free upgrade for PCs running Windows 7 or Windows 8. Microsoft has even backtracked on its original statement that Windows 10 will hit one billion devices by mid-2018, saying last month that Windows 10 likely won't in fact make that deadline.

An anonymous reader writes: Noah Smith argues that the field of economics frequently uses math in an unhealthy way. He says many economists don't use math as a tool to describe reality, but rather as an abstract foundation for whatever theory they've come up with. A possible solution to this, he says, is machine learning: "In other words, econ is now a rogue branch of applied math. Developed without access to good data, it evolved different scientific values and conventions. But this is changing fast, as information technology and the computer revolution have furnished economists with mountains of data. As a result, empirical analysis is coming to dominate econ. ... [Two economists pushing this change] stated that machine learning techniques emphasized causality less than traditional economic statistical techniques, or what's usually known as econometrics. In other words, machine learning is more about forecasting than about understanding the effects of policy. That would make the techniques less interesting to many economists, who are usually more concerned about giving policy recommendations than in making forecasts."

HughPickens.com writes: Kate Murphy writes at NYT about mathematician John Urschel whose latest contribution to the mathematical realm was a paper for the Journal of Computational Mathematics with the impressively esoteric title, "A Cascadic Multigrid Algorithm for Computing the Fiedler Vector of Graph Laplacians." "Believe me, I am aware that terms such as multigrid, Fiedler, and vector are not words that people use in their daily lives," says Urshel.

But as an offensive guard for the Baltimore Ravens, John Urschel regularly goes head to head with the top defensive players in the NFL and does his best to keep quarterback Joe Flacco out of harm's way. "I play because I love the game. I love hitting people," Urshel writes. "There's a rush you get when you go out on the field, lay everything on the line and physically dominate the player across from you. This is a feeling I'm (for lack of a better word) addicted to, and I'm hard-pressed to find anywhere else."

Urschel acknowledges that he has faced questions from NFL officials, journalists, fans and fellow mathematicians about why he runs the risk of potential brain injury from playing football when he has "a bright career ahead of me in mathematics" but doesn't feel able to quit. "When I go too long without physical contact I'm not a pleasant person to be around. This is why, every offseason, I train in kickboxing and wrestling in addition to my lifting, running and position-specific drill work."

Jon Brodkin writes "I’ve been a longtime fan of Brain Age. Mixed in among the standard-issue kill-everything-you-see/race/sports types of games that dominate gaming, Brain Age on the Nintendo DS always provided something unique, fun, and mentally stimulating. Doing math problems, counting syllables, recognizing patterns, and memorizing stuff was far more enjoyable than anyone would have expected in Brain Age: Train Your Brain in Minutes a Day, the game that kicked off the series seven years ago. Based on the research of neuroscientist Ryuta Kawashima, the exercises are designed to improve brain function—or at least give players the illusion that they’re getting smarter. Nintendo developed several sequels and spinoffs for the DS after that first game, and now a Nintendo 3DS-exclusive entry is here with Brain Age: Concentration Training, released as a download and physical media this past weekend for $30. Some improvements in gameplay are readily apparent. Handwriting recognition is significantly better. You play the game by holding the 3DS upright, rather than sideways like a book, and it works so well I wonder why previous Brain Age games used the wacky book-like layout at all." Read below to see what Jon thinks of the 3DS-exclusive version.

theodp writes "Slate's Ben Paynter looks into why Indian kids dominate the Scripps National Spelling Bee, and concludes it's because they have their own minor-league spelling bee circuit (having the discipline to spell 7,000 to 8,000 words a day probably helps too!). Indian-Americans make up about 1% of the US population, notes Paynter, but this year an estimated 11% of the competitors at Scripps will hail from regional contests run by the North South Foundation. The NSF competitions function as a kind of nerd Olympiad for Indian-Americans — there are separate divisions for math, science, vocabulary, geography, essay writing, and even public speaking — and a way to raise money for college scholarships for underprivileged students in India. BTW, Strollerderby has the scoop on Whatever Happened to the Spellbound Kids? (RIP, Ted Brigham)."

Coryoth writes "The widely held belief that there is disparity in the innate mathematical abilities of men and women has been steadily whittled down in recent years. The gender gap in basic mathematics skills closed some time ago, and recently the gap in high school mathematics has closed up as well, with as many girls as boys now taking high school calculus. Newsweek reports on a new study published in The Proceedings of the National Academy of Sciences that begins to lay to rest the remaining argument that it is at the highest levels of mathematics that the innate differences show. Certainly men dominate current academia, with 70% of mathematics Ph.D.s going to men; however that figure is down from 95% in the 1950s. Indeed, while there remain gaps in achievement between the genders, the study shows that not only are these gaps closing, but the size of the gap varies over differing cultures and correlates with the general degree of gender inequality in the culture (as defined by World Economic Forum measures). In all, this amounts to strong evidence that the differences in outcomes in mathematics between the genders is driven by sociocultural factors rather than innate differences in ability."

TechForensics writes "Outliers, by Malcolm Gladwell, is subtitled "the story of success." It is a book that purports to explain why some people succeed far more than others. It suggests that a success like Bill Gates is more attributable to external factors than anything within the man. Even his birth date turns out to play a role of profound importance in the success of Bill Gates and Microsoft Corporation." Look below for the rest of Leon's review.

anthemaniac writes "Computerized projections in sports are nothing new, but Bruce Bukiet of the New Jersey Institute of Technology has developed a model that seems to work pretty well. He projects how many games a Major League Baseball team will win by factoring in how each hitter ought to do against each pitcher in every game. His crystal ball says the Yankees will win 110 games this year, a pretty safe bet, many might agree. But he also projects all the divisional winners. He claims to be right more than wrong in five of the past six years."

A handful of updates, corrections and further thoughts on recent Slashdot stories follow; read on for updates on the Real-ID Act, Hollywood consultant math professor Jonathan Farley, the real first losers (and winners of the U.S. Open's Aibo League) at the 2005 Robo-Cup, and more. Details below.

Okay, here are Hans Reiser's answers to your questions about ReiserFS, starting and managing (and publicizing) a free software project, earning a living writing free software, and the good and bad points of being considered somewhat of a curmudgeon. As a free bonus, Hans adds a little insight into the politics of Linux kernel development, as in what gets accepted and what doesn't. Good stuff!

Kent M. Pitman has already given you his first 11 answers to the questions you asked him about Lisp, Scheme, the creation of programming standards, and much more -- below are his answers to another eight (starting with answer #12). Thanks again, Kent.

nomadic asks: " Reuters recently published an article about how American scientists, as well as foreign-born scientists who work in America, tend to dominate the Nobel prizes in science; it attributes this mostly to the fact that the U.S. government tends to invest more in science research than its foreign counterparts (the National Science Foundation funded 78 U.S. winners before they got their Nobels), and private and corporate entities contribute large amounts as well. The article talks about the scientific and economic culture of the U.S.; young scientists exist in an extremely competitive environment, where they are encouraged to challenge traditional authority. But it only touches on education a little. Now I've seen some truly venomous attacks on the U.S. educational system on /. by people in other countries, but this article implies while the U.S. is behind in science and math education in elementary and high school, 'there is something that happens on the college level'. Does the U.S. 'catch up' at the college level? I'll be honest, the exchange students I've met from more math and science-savvy educational systems have been well-educated, but not on some higher plane of thought that the media would lead one to believe. It seems commonly accepted (though I'm not sure I agree) that a high school graduate from most countries in Europe is on the average better-educated than one from the U.S. How about a college graduate? Graduate school? Is the U.S. dominance in science only about financial investment in the U.S., or can the U.S. university system take some credit?"

The Code Book remains a book good enough for another review, this time by the book-devouring stern. It's particularly interesting given the recent story that the code challenge it includes has been cracked. For another point of view, check out Jon Katz's review from a few months ago, nearly as enthusiastic. This book has already cost me several cups of coffee at BookPeople;), too!

Michael Crawford has written a feature on the recent trend towards integration and what it menas to us as the trend continue. Its a nice little summary piece of something that has been happening in the industry lately.