Argus Digest: test

Source contribution summary for this digest
Source	Type	Included	Scored	28d Digest Rate	28d Avg Score	28d Hotlist Hit	7d Article Age	28d Confidence
Guardian	news	1	10	10%	0.21	0%	11.0h	Stable
Hacker News	commentary	1	10	10%	0.24	0%	2.3d	Stable
Medium Artificial Intelligence (keyword)	commentary	1	10	6%	0.10	0%	0.8h	Stable
MyFT	news	1	10	6%	0.05	0%	4.2h	Stable
Reddit ArtistHate	news	1	10	Collecting data	Collecting data	Collecting data	No recent data	Collecting
Zero Hedge	commentary	1	10	3%	0.01	0%	9.6h	Stable
arXiv CompSci CL	research	1	10	~6%	~0.10	~0%	3.8h	Low sample
arXiv CompSci ML	research	1	10	~9%	~0.09	~0%	3.8h	Low sample
TechCrunch	news	1	8	10%	0.11	0%	8.4h	Stable
Seeking Alpha News	commentary	1	7	2%	0.00	0%	1.3h	Stable
Ars Technical All News	news	1	6	11%	0.12	0%	8.0h	Stable
WSJ Tech	news	1	5	8%	0.09	0%	7.5h	Stable
The Verge	news	1	2	12%	0.16	0%	7.8h	Stable
MIT Research General	research	1	1	Collecting data	Collecting data	Collecting data	4.9h	Collecting
Wired AI News	news	1	1	~16%	~0.16	~0%	9.5h	Low sample
Bloomberg Markets	news	0	10	3%	0.02	0%	5.0h	Stable
Bogleheads forum	news	0	10	5%	0.05	0%	1.0h	Stable
NYT front page	news	0	10	7%	0.07	0%	8.9h	Stable
Reddit R/DayTrading	news	0	10	Collecting data	Collecting data	Collecting data	No recent data	Collecting
WSJ US Business	news	0	10	5%	0.06	0%	10.0h	Stable
Medium AI (keyword)	commentary	0	8	6%	0.10	0%	0.8h	Stable
India Times	news	0	6	1%	0.01	0%	2.6h	Stable
The Atlantic	news	0	4	10%	0.10	0%	8.9h	Stable
FT Alphaville	news	0	3	~13%	~0.12	~0%	5.3h	Low sample
Reddit R/Ethereum	news	0	2	~9%	~0.08	~0%	2.8h	Low sample
Atlas Obscura	commentary	0	1	~31%	~0.65	~0%	11.7h	Low sample
Closing Line Substack	news	0	1	Collecting data	Collecting data	Collecting data	8.8h	Collecting
Economist: Business	news	0	1	Collecting data	Collecting data	Collecting data	3.6h	Collecting
Economist: Leaders	news	0	1	Collecting data	Collecting data	Collecting data	10.1h	Collecting
Economist: United States	news	0	1	Collecting data	Collecting data	Collecting data	7.6h	Collecting
El Reg Offbeat	news	0	1	Collecting data	Collecting data	Collecting data	5.8h	Collecting
Futurism	news	0	1	12%	0.14	0%	5.7h	Stable
Grumpy Economist (Cochrane)	commentary	0	1	Collecting data	Collecting data	Collecting data	9.7h	Collecting
Latent Space	commentary	0	1	Collecting data	Collecting data	Collecting data	9.4h	Collecting
NYT Economy	news	0	1	Collecting data	Collecting data	Collecting data	7.8h	Collecting
Noahpinion	commentary	0	1	Collecting data	Collecting data	Collecting data	6.1h	Collecting
Outside Law School Scam - Comments	commentary	0	1	Collecting data	Collecting data	Collecting data	9.4h	Collecting
Silver Bulletin	commentary	0	1	Collecting data	Collecting data	Collecting data	2.1h	Collecting
Venture Beat	commentary	0	1	~0%	~0.03	~0%	6.9h	Low sample
WSJ Social Economy	news	0	1	5%	0.01	0%	6.1h	Stable
ZD Net	news	0	1	9%	0.13	0%	8.0h	Stable

I Saw the Future of AI in a Robot That Can Learn on the Spot

Wired AI News | positive | Published: 15:30 Aug 19, 2026 (Eastern)

Wired AI News reports on a visit to Generalist AI, a Cambridge, Massachusetts-based startup founded by Pete Florence, Andrew Barry, and Andy Zeng, which is developing robot arms capable of learning tasks on the spot from short instructional videos without specific prior training. During demonstrations, the robots improvised when tools were missing—such as using a dustpan as a brush or switching grippers to better retrieve banknotes—showcasing a form of physical intelligence inspired by human toddlers. The company builds its models from scratch, utilizing a large scale of physical interaction data collected through human-worn, camera-equipped special grippers. While roboticists note the company's strong execution and potential for commercial deployment, the robots currently achieve a success rate of only about 59 percent and reliability remains a challenge.

Keywords: robotics, Generalist AI, banana, robotic arm, whimsical technology

Why I, a Retired Botanist, Now Code With AI

Medium Artificial Intelligence (keyword) | positive | Published: 02:55 Aug 20, 2026 (Eastern)

A retired botanist with a forty-year career studying cyanobacteria and a late-in-life review book on algal chromosomes discusses their experience coding with an AI coding agent.

Keywords: botanist, artificial intelligence, coding, cyanobacteria, retirement

A faster way to calculate the day of the week

Hacker News | neutral | Published: 17:20 Aug 16, 2026 (Eastern)

Ben Joffe's article explores methods for efficiently calculating the day of the week from a day-count ("rata-die") integer. The post presents several fast 32-bit and 64-bit algorithms tuned for different throughput and latency requirements across platforms like x86 and ARM, improving on existing techniques by Howard Hinnant and Cassio Neri. Additionally, the author introduces novel fast modulus formulas for divisors such as 24 and 60, applicable to timekeeping and date library optimizations.

Keywords: calendar, math, mental calculation, day of the week

Gwyneth Paltrow allegedly set to throw dinner in honor of Sam Altman

TechCrunch | neutral | Published: 16:04 Aug 19, 2026 (Eastern)

According to Puck, actress, Goop founder, and Kinship Ventures investor Gwyneth Paltrow is reportedly hosting a private, off-the-record dinner at her Hamptons home on August 29 in honor of OpenAI CEO Sam Altman. Kinship Ventures is an investor in OpenAI and other AI companies, and Paltrow has previously discussed using AI tools to run her business, while Goop Kitchen leased space in Anthropic's San Francisco headquarters.

Keywords: Gwyneth Paltrow, Sam Altman, Kinship Ventures, dinner

Urge Boom Studios to accept fan Ranger designs for comics

Reddit ArtistHate | neutral | Published: 01:11 Aug 17, 2026 (Eastern)

A Reddit post from the ArtistHate community encourages readers to urge Boom Studios to accept fan-created Ranger designs for comics.

Keywords: Boom Studios, Power Rangers, comic books, fan art, petition

Meet The Pro-Trump Billionaire Who Paid $40 Million For Fugly Ferrari EV

Zero Hedge | neutral | Published: 18:00 Aug 19, 2026 (Eastern)

Florida billionaire and optometrist Dr. Herbert Wertheim, known for his support of Donald Trump, was identified as the buyer who paid $40 million for a unique Ferrari EV model called 'Chassis 0' or the 'Luce.' Wertheim, who has an estimated net worth of $4.6 billion, plans to use the vehicle for charitable fundraising rather than as a personal car. The article also notes his previous high-value auction purchases, regular visits to Mar-a-Lago, and a $2 million winning bid at a charity gala for a private White House visit with Trump.

Keywords: Ferrari EV, billionaire, auction, Herbert Wertheim, cars

Teenage ace Rafael Jódar endures defeat by Cobolli but his star is rising

Guardian | mixed | Subscription | Published: 17:33 Aug 19, 2026 (Eastern)

Nineteen-year-old tennis player Rafael Jódar was defeated by Flavio Cobolli in the fourth round of the Cincinnati Open, losing 4-6, 7-6 (3), 6-3 after coming close to victory. Despite the loss, the article details Jódar's rapid rise during his first full professional season, including winning his first ATP title in Marrakech, reaching a grand slam quarter-final at Roland Garros, and climbing to a ranking of No 11. The piece also discusses his playing style and shot-making ability, alongside criticisms regarding his on-court demeanor, such as taking medical timeouts, frequent shoelace-related delays, and impatience with ballkids and volunteers.

Keywords: Rafael Jódar, Flavio Cobolli, Cincinnati Open, tennis, ATP Tour

L\'evy Attention: Single-Pass Predictive Uncertainty for Continuous-Time Attention

arXiv CompSci ML | neutral | Published: 00:00 Aug 20, 2026 (Eastern)

The article introduces Lévy Attention, a cross-attention operator designed for continuous-time attention in deep models for irregularly-sampled time series. The method uses a stochastic formulation based on an inhomogeneous Poisson random measure to compute predictions and provide a closed-form measure of predictive uncertainty in a single pass without extra trained heads. In experiments, the operator achieves comparable accuracy to controls while providing uncertainty signals that outperform multi-pass Monte Carlo dropout and scale calibrated Gaussians.

Keywords: Lévy Attention, machine learning, predictive uncertainty, time series, stochastic formulation, artificial intelligence

The Hidden Debt That Apple Owes to the CIA

WSJ Tech | neutral | Subscription | Published: 19:00 Aug 19, 2026 (Eastern)

According to a WSJ Tech article, Apple and the iPhone owe a hidden debt to the CIA, as CIA funding helped keep Steve Jobs's company NeXT afloat in the 1980s, enabling his eventual return to Apple and providing the foundational operating code for the iPhone.

Keywords: Apple, CIA, Steve Jobs, NeXT, technology history

Apple still eyes 2027 launch for AI-powered AirPods despite video leak - report

Seeking Alpha News | neutral | Published: 03:06 Aug 20, 2026 (Eastern)

According to a report from Seeking Alpha News, Apple is still targeting a 2027 launch for its AI-powered AirPods, despite a recent video leak.

Keywords: Apple, AirPods, AI, technology, leak

NASA calls off mission to rescue Swift gamma-ray observatory

Ars Technical All News | negative | Published: 20:18 Aug 19, 2026 (Eastern)

NASA and Katalyst Space Technologies have canceled a robotic mission to rescue the Swift gamma-ray telescope before it reenters Earth's atmosphere. The rescue satellite, named Link, launched on July 3 under a $30 million contract awarded to startup Katalyst on a nine-month schedule. Although the spacecraft experienced ongoing attitude control issues after two of its three reaction wheels and cold gas thrusters stopped working in late July, Katalyst stated that the satellite remains operational and will be used to extract remaining technical value and potentially perform a close-in navigation demonstration.

Keywords: NASA, Swift gamma-ray observatory, space mission, satellite reentry

Blood clot discovery upends consensus about how wounds heal

MyFT | neutral | Subscription | Published: 00:00 Aug 20, 2026 (Eastern)

Scientists have discovered that proteins stack up in layers like sheets of paper to initiate the formation of blood clots, altering the consensus on how wounds heal.

Keywords: blood clot, wound healing, science, proteins, medical discovery

MIT engineers design a better controller for operating construction diggers

MIT Research General | neutral | Published: 00:00 Aug 20, 2026 (Eastern)

MIT engineers, in collaboration with Sumitomo Heavy Industries, have designed a new training and control interface called the "World-Space Interface" (WSI) to shorten the learning curve for operating construction excavators. Instead of using traditional, non-intuitive joysticks, the WSI features a miniature mechanical arm and bucket that trainees grasp to mimic the movements of an excavator, which are mirrored in a virtual environment across an immersive six-screen display. In experiments comparing the WSI to a traditional joystick simulator over a seven-day training period, researchers found that novices using the WSI performed as well as experts from the start. The researchers plan to further improve the system by adding haptic feedback to simulate physical forces, and envision the interface being used for faster operator training, inside excavator cabs as an exoskeleton-like arm, or for remote tele-operation in unsafe environments.

Keywords: MIT, engineers, construction diggers, excavator, controller

Does giving a camera wings dodge the FCC’s drone ban?

The Verge | neutral | Published: 18:35 Aug 19, 2026 (Eastern)

HoverAir is attempting to bypass the US government's December 2025 ban on future foreign drones with a new device, the HoverAir Versa, according to an article from The Verge.

Keywords: HoverAir, drones, FCC ban, gadgets, US government

Corrections of Zipf's and Heaps' Laws Derived from Hapax Rate Models

arXiv CompSci CL | neutral | Published: 00:00 Aug 20, 2026 (Eastern)

The research article titled "Corrections of Zipf's and Heaps' Laws Derived from Hapax Rate Models" introduces corrections to Zipf's and Heaps' laws derived from systematic models of the proportion of hapaxes (words occurring once). The derivation is based on the standard urn model, which assumes shorter text marginal frequency distributions mimic blind sampling from a longer text, and an assumption that the hapax rate is a simple function of text length. Four functions are evaluated—the constant, cancelation, linear, and logistic models—with the logistic model yielding the best fit for a sample of 14 English texts. Additionally, the paper discusses mixture models reflecting two-regime vocabularies for larger corpora.

Keywords: Zipf's Law, Heaps' Law, Hapax Legomena, Linguistics, Statistical Models