In partnership with developers, game marketplace GOG (Good Old Games) has launched a new website called FreedomtoBuy.games that’ll let you download select “adult” games for free. GOG believes the website takes a stand “against the quiet erasure of creative works from digital shelves,” a response of sorts to recent decisions from Steam and Itch to delist certain violent and sexuality-explicit games from their respective platforms.
GOG is currently offering 13 games for free for the next 48 hours, some with well-known scandals and others that seem to fall into the “NSFW visual novel” bucket that makes up the majority of sexually-explicit games on digital storefronts. The titles available to download are:
POSTAL 2, a graphically violent open-world game, is a notable inclusion because it was banned in New Zealand in 2004 and delisted from the German version of Steam in 2016. HuniePop, one of several “adult-only” games Twitch streamers are explicitly forbidden to stream, makes sense on the list, too. GOG has made a concerted effort to preserve games of all types, including maintaining them so that they run on current hardware. The point of making these games available to download is as much about preservation as it is about highlighting how apparently easy it is to pressure digital storefronts to remove content, though.
Valve’s decision to delist titles from Steam was chalked up to a new rule that requires games to abide by the standards set by the payment processors that work with Steam. Itch offered a similar explanation for the delistings on its storefront, pointing to pressure payment processors were receiving from a nonprofit called Collective Shout. In the process, indie games like VILE: Exhumed have been delisted, primarily for depicting things that might make a certain group of people uncomfortable.
Itch, for its part, seems to be trying to bring back as many games to its storefront as it can. The platform is restoring free NSFW games, and says it’s still in talks with its payment partners about restoring paid games to its storefront.
As they spin, comet nuclei often show periodic variations of their brightness over timescales of order hours. Astronomers expect that to apply also to the new interstellar object 3I/ATLAS, since it was classified by the Minor Planet Center as a comet.
The first interstellar object, 1I/`Oumuamua was spinning with a period of about 8 hours, but its spin did not evolve as expected from a comet with its measured non-gravitational acceleration. Even more paradoxically, deep observations by the Spitzer space telescope did not find carbon-based molecules or dust in a cometary tail around 1I/`Oumuamua. As in Hans-Christian Andersen’s folktale, this “emperor” had no visible clothes but the astronomers pretending to be the adults in the room insist in a recent paper that 1I/’Oumuamua is a dark comet with invisible clothes.
As of now, there is no spectroscopic evidence for atomic or molecular gas around 3I/ATLAS either (see the papers here, here and here). In addition, data from the new Rubin Observatory on 3I/ATLAS shows no variability on hourly timescales. Why isn’t the flux from 3I/ATLAS changing periodically? Either 3I/ATLAS is not rotating or the rotation of its nucleus is hidden from view deep inside a thick cloud of dust.
Another interesting fact gleaned from images of 3I/ATLAS is that it has a leading glowing halo rather than a trailing tail. Why is there a glow ahead of 3I/ATLAS? One possible explanation is that the object does not spin and so its dayside is hot and its nightside is cold. As a result, any ice that accumulated on its surface during its freezing interstellar travel evaporated only at the front of the object facing the Sun. This would mean that the object’s surface is not hidden behind a veil of dust and so its diameter must be 10–20 kilometers to explain its brightness. This brings back the argument from my published paper that 3I/ATLAS cannot be a 10–20 kilometer rock because the available rocky material in interstellar space can only deliver a 10–20 kilometer rock to the survey volume of the ATLAS survey telescope once per 10,000 years rather than once per 10 years as required. If 3I/ATLAS is not a 10–20 kilometers rock, then what is the origin of this object?
One way to avoid the mass budget problem is by delivering 3I/ATLAS to the inner solar system through a fine-tuned trajectory. As it turns out, the trajectory of 3I/ATLAS has indeed rare features. It lines up with the orbital plane of the inner planets (with a likelihood of 0.2%) and comes unusually close to Jupiter, Mars and Venus (with a likelihood of 0.005%). Is 3I/ATLAS alien technology? This was the title of my second paper on 3I/ATLAS, in collaboration with Adam Hibberd and Adam Crowl from the Initiative for Interstellar Studies.
In recent days, I was independently asked by dozens of podcasters and reporters for my opinion on the probability that 3I/ATLAS is an object under intelligent control. I replied that it is difficult to assign a probability to the technological origin. It is natural to expect 3I/ATLAS to be a comet with a compact nucleus surrounded by a large dust cloud. Yet, 3I/ATLAS follows an exceptionally rare trajectory. And so, the final verdict depends on the significance assigned to the trajectory’s peculiarities. We will have a better assessment of this probability within a month or two, once 3I/ATLAS gets brighter and easier to observe as it approaches the Sun.
Given that the NSF-DOE Rubin Observatory is expected to find a new interstellar object every few months in the coming decade, I suggest that the time is ripe to establish a “Loeb scale” for the risk from interstellar objects similar to the Richter scale for earthquakes. The zero point of that scale would correspond to a verified comet or asteroid of a natural origin and a value of 10 would mean a verified technological object, possibly propelled by an engine which powers artificial lights. A team of policy experts, in consultation with scientists, will decide how to respond to an alien threat. And a team of psychologists will suggest how to communicate the message to the public without creating panic and a meltdown of world economies.
It is impossible to resolve 3I/ATLAS with telescopes on Earth, but we can get a better look of it with a camera on a spacecraft that gets close to its path. In my latest paper with Adam Hibberd and Adam Crowl, I proposed using the Juno spacecraft around Jupiter with this goal in mind. The Juno mission is currently scheduled to end its life in mid-September 2025. Instead of giving the spacecraft a push towards Jupiter’s atmosphere where it will burn up to ashes, the paper proposes giving it extra scientific life by instructing Juno’s engine to give it a push away from Jupiter — so that by mid-March 2026 it will intercept 3I/ATLAS at its closest approach to Jupiter. If the available fuel is insufficient for an intercept, even a mild burn of the engine in mid-September 2025 could bring Juno closer to the path of 3I ATLAS by mid-March 2026.
On July 29, 2025, I received a call from a member of the U.S. House of Representatives, the brilliant Anna Paulina Luna, who asked me to brief her about 3I/ATLAS. At the end of our conversation, she kindly offered to send a letter that encourages the leadership of NASA to use Juno as a probe of 3I/ATLAS.
A close encounter between Juno and 3I/ATLAS will provide a once-in-a-lifetime opportunity for human-made technologies to come within a short distance from a large interstellar object. This could actually happen in 2026 rather than appear in a Hollywood science-fiction script such as Stanley Kubrick’s film “2001: A Space Odyssey.” The encounter will also be an unprecedented public outreach opportunity for NASA to engage with the viral interest of followers of Joe Rogan and other podcasters in the nature of 3I/ATLAS.
As a token of my appreciation and gratitude for the vision of the Honorable Congresswoman Anna Paulina Luna, I attach below the letter that she sent to NASA today (also accessible here). The letter encourages NASA’s leadership to conduct a study of the feasibility of using Juno as a probe of 3I/ATLAS.
Zoom image will be displayed
Zoom image will be displayed
If aliens are watching Earth right now, they will likely raise humanity’s intelligence ranking among the stars after reading Congresswoman Luna’s inspiring letter.
Here’s hoping that a visit by an interstellar messenger will usher in a peaceful messianic era for all earthlings.
ABOUT THE AUTHOR
Zoom image will be displayed
(Image Credit: Chris Michel, National Academy of Sciences, 2023)
Avi Loeb is the head of the Galileo Project, founding director of Harvard University’s — Black Hole Initiative, director of the Institute for Theory and Computation at the Harvard-Smithsonian Center for Astrophysics, and the former chair of the astronomy department at Harvard University (2011–2020). He is a former member of the President’s Council of Advisors on Science and Technology and a former chair of the Board on Physics and Astronomy of the National Academies. He is the bestselling author of “Extraterrestrial:The First Sign of Intelligent Life Beyond Earth” and a co-author of the textbook “Life in the Cosmos”, both published in 2021. The paperback edition of his new book, titled “Interstellar”, was published in August 2024.
Meteor activity kicks into high gear in August as seen from the northern hemisphere. The main reason for all this activity is the Perseid shower that peaks on August 13th. This shower is active most of the month and remains above the level of the sporadic background for a week centered on August 13th. The sporadic activity is also near maximum as seen from the northern hemisphere and is now more than double the rates from just three months ago. As seen from the southern hemisphere, meteor rates are still decent but falling rapidly. The sporadic rates continue their downward slide plus the Perseid radiant does not rise high into the sky as seen in the southern hemisphere. Therefore, rates for the Perseids are greatly reduced when compared to those seen from the northern hemisphere.
During this period, the moon waxes from half-illuminated to nearly full during this period. This weekend the moon will set during the early morning hours allowing a few hours of observations under dark conditions between moon set and dawn. This window of opportunity shrinks with each passing night until mid-week when it lies above the horizon all night long. The estimated total hourly rates for evening observers this weekend should be near 4 as seen from mid-northern latitudes (45N) and 3 as seen from tropical southern locations (25S). For morning observers, the estimated total hourly rates should be near 21 as seen from mid-northern latitudes (45N) and 18 as seen from tropical southern locations (25S). The actual rates seen will also depend on factors such as personal light and motion perception, local weather conditions, alertness, and experience in watching meteor activity. Evening rates during this period are reduced due to moonlight. Note that the hourly rates listed below are estimates as viewed from dark sky sites away from urban light sources. Observers viewing from urban areas will see less activity as only the brighter meteors will be visible from such locations.
The radiant (the area of the sky where meteors appear to shoot from) positions and rates listed below are exact for Saturday night/Sunday morning August 2/3. These positions do not change greatly day to day so the listed positions may be used during this entire period. Most star atlases (available online and at bookstores and planetariums) will provide maps with grid lines of the celestial coordinates so that you may find out exactly where these positions are located in the sky. I have also included charts of the sky that display the radiant positions for evening, midnight, and morning. The center of each chart is the sky directly overhead at the appropriate hour. These charts are oriented for facing south but can be used for any direction by rotating the charts to the desired direction. A planisphere or computer planetarium program is also useful in showing the sky at any time of night on any date of the year. Activity from each radiant is best seen when it is positioned highest in the sky (culmination), either due north or south along the meridian, depending on your latitude. Radiants that rise after midnight will not reach their highest point in the sky until daylight. For these radiants, it is best to view them during the last few hours before dawn. It must be remembered that meteor activity is rarely seen at its radiant position. Rather they shoot outwards from the radiant, so it is best to center your field of view so that the radiant lies toward the edge and not the center. Viewing there will allow you to easily trace the path of each meteor back to the radiant (if it is a shower member) or in another direction if it is sporadic. Meteor activity is not seen from radiants that are located far below the horizon. The positions below are listed in a west to east manner in order of right ascension (celestial longitude). The positions listed first are located further west therefore are accessible earlier in the night while those listed further down the list rise later in the night.
Radiant Positions at 23:00 Local Summer Time
Radiant Positions at 01:00 Local Summer Time
Radiant Positions at 03:00 Local Summer Time
These sources of meteoric activity are expected to be active this week
.
The kappa Cygnids (KCG) are active from July 23-August 27, with maximum occurring on August 14th. The radiant is currently located at 18:38 (280) +44. This area of the sky is located in northern Lyra, 5 degrees north of the zero-magnitude star known as Vega (alpha Lyrae). To best see these meteors face northward near 2300 Local Summer Time (LST) when it lies on the meridian and is located highest in the sky. With a high northern declination, these meteors are difficult to view from the southern hemisphere. Expected hourly rates this week are less than 1 no matter your location. With an entry velocity of 23 km/sec., the average meteor from this source would be of medium-slow velocity.
The alpha Capricornids (CAP) are active from July 7 through August 13, peaking on July 30th. The radiant is currently located at 20:44 (311) -09. This position lies southeastern Aquila, 2 degrees west of the 4th magnitude star known as delta Aquarii. Current rates are expected to be near 1 per hour no matter your location. These meteors are best seen near 01:00 LST, when the radiant lies highest in the northern sky. With an entry velocity of 21 km/sec., the average meteor from this source would be of medium-slow velocity.
The large Anthelion (ANT) radiant is currently centered at 21:32 (323) -13. This position lies in central Capricornus, 4 degrees northeast of the 4th magnitude star known as iota Capricornii. This radiant is best placed near 02:00 LST when it lies on the meridian and is nearly overhead. Rates at this time should be near 1 per hour as seen from the northern hemisphere and 2 per hour as seen from south of the equator. With an entry velocity of 30 km/sec., the average Anthelion meteor would be of medium-slow velocity.
The Southern delta Aquariids (SDA) are active from July 19 through August 13 with maximum activity occurring on July 30. The radiant is currently located at 22:52 (343) -15. This area of the sky is located in southern Aquarius, 1 degree west of the 3rd magnitude star known as Skat (delta Aquarii). This radiant is best placed near 0300 LST, when it lies on the meridian and is nearly overhead. Hourly rates at this time should be near 1 as seen from the northern hemisphere and near 2 as seen from south of the equator. With an entry velocity of 41 km/sec., the average meteor from this source would be of medium velocity.
The Perseids (PER) are active from July 17 through August 29, with maximum activity occurring on August 13. The radiant is currently located at 02:16 (034) +55. This position lies in northwestern Perseus, 5 degrees southwest of the 4th magnitude star known as Miram (eta Persei). This area of the sky is best placed for viewing during the last dark hour before dawn when it lies highest in the northeastern sky. Maximum activity is not until August 12th so current rates are expected to be near 5 as seen from the northern hemisphere and 2 as seen from south of the equator. With an entry velocity of 59 km/sec., the average meteor from this source would be of swift velocity. Viewers in the southern hemisphere have a limited view of this shower as the radiant only rises just before dawn.
The eta Eridanids (ERI) are active from a radiant near 02:34 (038) -14. This position lies in southeastern Cetus, 3 degrees west of the 4th magnitude star known as pi Ceti. This source is active until September 10th, with maximum activity occurring on August 4th. Current rates are expected to be near 1 per hour no matter your location. These meteors are best seen during the last dark hour prior to dawn when the radiant lies highest above the southeastern horizon in a dark sky. With an entry velocity of 64 km/sec., the average meteor from this source would be of swift speed.
Sporadic meteors are those meteors that cannot be associated with any known meteor shower. All meteor showers are evolving and disperse over time to the point where they are no longer recognizable. Away from the peaks of the major annual showers, these sporadic meteors make up the bulk of the activity seen each night. As seen from the mid-northern hemisphere (45N) one would expect to see during this period approximately 12 sporadic meteors per hour during the last hour before dawn as seen from rural observing sites. Evening rates would be near 3 per hour. As seen from the tropical southern latitudes (25S), morning rates would be near 9 per hour as seen from rural observing sites and 2 per hour during the evening hours. Locations between these two extremes would see activity between these listed figures. Evening rates are reduced due to moonlight.
The list below offers information in tabular form. Rates and positions in the table are exact for Saturday night/Sunday morning.
SHOWER
DATE OF MAXIMUM ACTIVITY
CELESTIAL POSITION
ENTRY VELOCITY
CULMINATION
HOURLY RATE
CLASS
RA (RA in Deg.) DEC
Km/Sec
Local Summer Time
North-South
kappa Cygnids (KCG)
Aug 16
18:38 (280) +44
23
23:00
<1 – <1
II
alpha Capricornids (CAP)
Jul 31
20:44 (311) -09
23
01:00
1 – 1
II
Anthelion (ANT)
–
21:32 (323) -13
30
02:00
1 – 2
II
Southern delta Aquariids (SDA)
Jul 31
22:52 (343) -15
41
03:00
1 – 2
I
Perseids (PER)
Aug 12
02:22 (036) +56
59
06:00
5 – 2
I
eta Eridanids (ERI)
Aug 07
02:34 (038) -14
64
06:00
1 – 1
II
Class Explanation: A scale to group meteor showers by their intensity:
Class I: the strongest annual showers with Zenith Hourly Rates normally ten or better.
Class II: reliable minor showers with ZHR’s normally two to ten.
Class III: showers that do not provide annual activity. These showers are rarely active yet have the potential to produce a major display on occasion.
Class IV: weak minor showers with ZHR’s rarely exceeding two. The study of these showers is best left to experienced observers who use plotting and angular velocity estimates to determine shower association. These weak showers are also good targets for video and photographic work. Observers with less experience are urged to limit their shower associations to showers with a rating of I to III.
Get our favorite new stories right to your inbox every week
This article is an installment of Future Explored, a monthly guide to world-changing technology. You can get stories like this one straight to your inbox every month by subscribing above.
It’s 2040. Drug discovery is booming, thanks to virtual cells. These AI-powered models of living cells have become indispensable tools in biomedical research, helping scientists test treatments in silico before they ever reach a lab — saving time, money, and lives.
Virtual Cells
Artificial intelligence is quickly becoming biology’s most powerful microscope.
Top research centers are using the tech to develop virtual cells, AI-based simulations of the core building blocks of all lifeforms — and it’s hard to overstate the impact their models could have on the world of health.
“The vision that we can really understand everything about a cell — from its molecular structure to its function to how cells interact and operate in living organisms to how they respond and react to any intervention — will go a long way to helping us cure, prevent, and manage disease,” said Patricia Brennan, VP of Science Technology and General Manager for Science at the Chan Zuckerberg Initiative (CZI).
To find out how we got here — and where we’re going — this month’s Future Explored is taking a close look at virtual cells: what they are, who’s making them, and how they could shape the future of medicine.
Where we’ve been
Where we’re going (maybe)
The first virtual cell was built over a decade ago — so why is interest in them surging now? The answer lies in fundamental differences in how today’s models work compared to ones like what the Stanford team created in 2012.
Their virtual cell simulated all of the molecular processes of Mycoplasma genitalium (Mgen) by breaking them into 28 modules, each governed by its own set of mathematical equations and biological rules — a simplified example might be, “If Gene A is expressed, Gene B will be activated.”
The team developed those instructions by painstakingly digging through more than 900 papers, books, and databases, essentially distilling all of the scientific community’s knowledge of how Mgen functions into 28 algorithms. Once complete, their model could simulate the entire Mgen lifecycle, from birth to division, in about 10 hours — roughly the same amount of time an actual Mgen cell takes to divide.
Simply being able to observe this process was valuable, but the real utility of the virtual cell was that researchers could experiment on it. They could knock out a gene with keystrokes instead of CRISPR and then run the simulation to predict how its loss would affect Mgen. If they saw something interesting, they could then spend time on a lab experiment.
“If you use a model to guide your experiments, you’re going to discover things faster,” said study leader Markus Covert in 2012. “We’ve shown that time and time again.”
Most of the time, the virtual cell’s predictions would match the results of real-world experiments. When they didn’t, the discrepancies usually involved genes that were poorly understood in scientific literature. That made sense: The model was limited by its programming. If scientists wanted to improve it, they’d need to update their algorithms.
“If you use a model to guide your experiments, you’re going to discover things faster.”
Markus Covert
In the years following the Stanford breakthrough, other groups made their own virtual cells using the same method: distill the known literature into rule-based algorithms.
Today’s virtual cells, however, are built on artificial intelligence, usually a specific kind of model called a transformer. Google researchers first proposed this AI architecture in 2017, and it’s the basis for many of today’s best generative AIs, including ChatGPT.
Transformer-based AIs learn to spot relationships between tokens (small units of data) by training on huge datasets. Once trained, they can then generate new content by predicting the most likely next token in a sequence.
For ChatGPT, for example, tokens are words or parts of words. The huge training dataset was the internet, and once trained, ChatGPT was able to generate text by predicting the most likely token to come next in its response, over and over again, based on the ones that came before it.
One of the most remarkable things about transformer-based AIs is that they can output content that isn’t included in their training data. An AI image generator, for example, can output a photorealistic picture of a cat made of spaghetti even if it wasn’t explicitly shown what that should look like.
“The goal is for virtual cell models to serve as digital twins or computational stand-ins for experimental systems.”
Marinka Zitnik
Researchers are now building virtual cells on the transformer architecture — and the results are remarkable.
CZI’s TranscriptFormer model, for example, was trained on datasets containing images, RNA sequences, and other biological data from 112 million cells. These were drawn from 12 different species across 1.5 billion years of evolution. A researcher can now prompt it with data from a cell they’re studying, and the AI can predict its cell type, infection status, and more — even if the cell comes from a species that wasn’t included in the model’s training data.
“We just trained it on natural variability, but this natural variability follows the tree of life: It has a lot of structure. There’s a lot of interesting stuff happening there,” said Theofanis Karaletsos, Senior Director of AI at CZI. “The model actually becomes extremely rich and extremely performant on doing all kinds of tasks.”
CZI is now offering researchers early access to a one-stop platform that includes several virtual cell models, including TranscriptFormer, and the datasets used to train them. These models have specific use cases — CZI’s GREmLN model, for example, predicts how genes work together — but the vision is to develop models that can simulate more complex cellular behavior.
“Looking forward, the goal is for virtual cell models to serve as digital twins or computational stand-ins for experimental systems,” said Marinka Zitnik, a CZI collaborator and assistant professor at Harvard. “For example, a validated virtual cell could simulate the outcome of a drug or genetic intervention in silico, potentially reducing the need for animal experiments or guiding the design of laboratory studies.”
“You ask the model, ‘What perturbations do I need to make to move this cell from this diseased state to this healthy state?’”
David Burke
Biomedical research nonprofit Arc Institute is building virtual cells, too. It recently opened access to its first model, STATE, which was trained on observational data from nearly 170 million cells and perturbational data from over 100 million cells. Perturbational data captures how a cell reacts when its normal function is disrupted by a drug, genetic edit, or some other external stimulus.
Researchers input a cell’s transcriptome — the full collection of genes being actively expressed at a given moment — and a proposed perturbation, and STATE predicts how the cell’s gene expression patterns are likely to change. This can give scientists a way to test the potential impact of disease treatments without having to actually perform experiments.
By running this process in reverse, STATE can even point researchers toward promising interventions they hadn’t considered.
“You take a cell that’s in a diseased state — like maybe it’s got an Alzheimer’s disease transcriptomic profile — and one in a healthy state, and then you ask the model, ‘What perturbations do I need to make to move this cell from this diseased state to this healthy state?’” said David Burke, Arc Institute’s CTO.
STATE’s predictions range in accuracy from 40% to 60%, depending on the type of perturbation, according to Burke. He thinks 75% would be good enough that biologists could start relying on the AI’s predictions without having to run experiments in the wet lab.
“That might seem a bit low,” said Burke, “but when you look at all the different datasets from different labs, the concordance between them is only about 75% because single-cell sequencing and perturbation screens are very noisy, so that’s our target.”
“We will need a lot more data.”
Patricia Brennan
A transformer model is only as good as the quality and quantity of its training data, so if we want to improve today’s virtual cells, we’re going to need to improve our datasets.
“While the scale of the data sets has been growing over the last number of years, we will need a lot more data,” said CZI’s Brennan, who noted that much of the data we already have wasn’t necessarily collected with the training of virtual cells in mind, which complicates its use as training material.
To help close the data gap as quickly as possible, CZI launched the open-source Billion Cells Project (BCP) in February. The goal of the initiative is to quickly and cost-effectively generate a public dataset containing one billion cells through partnerships with scientists and the developers of cutting-edge cell analysis technologies.
“Traditional data generation pipelines can take three to four years, whereas BCP is compressing that timeline to months,” said Bailey Marshall, Senior Program Associate, Single-Cell Biology at CZI.
“One of the project’s most important innovations is in interoperability,” she added. “By aligning from the outset on standardized kits, protocols, and technologies across the BCP…data from diverse tissues, species, and modalities can be easily integrated. This allows researchers and AI developers to train models that are consistent, reproducible, and broadly applicable.”
“How far can we go? That’s an open scientific question.”
David Burke
So, in the short term, the scientific community knows it needs more data to make its virtual cells more robust, but the extent to which it can improve them is still unknown.
With enough high-quality data, is it possible to make a single model that accurately predicts everything that will happen in a cell when it’s perturbed in every way possible? How about a model that takes into account how cells work in context with one another? Can we create virtual models of entire tissues, organs, or even bodies?
If so, it would mark a turning point in biology — moving from a science that observes life to one that can simulate and predict it. The consequences for medicine, longevity, and our understanding of health would be transformative.
“How far can we go? That’s an open scientific question,” said Burke.
We’d love to hear from you! If you have a comment about this article or if you have a tip for a future Freethink story, please email us at [email protected].
Subscribe to Freethink on Substack for free
Get our favorite new stories right to your inbox every week
Anthropic revoked OpenAI’s API access to its models on Tuesday, multiple sources familiar with the matter tell WIRED. OpenAI was informed that its access was cut off due to violating the terms of service.
“Claude Code has become the go-to choice for coders everywhere and so it was no surprise to learn OpenAI’s own technical staff were also using our coding tools ahead of the launch of GPT-5,” Anthropic spokesperson Christopher Nulty said in a statement to WIRED. “Unfortunately, this is a direct violation of our terms of service.”
According to Anthropic’s commercial terms of service, customers are barred from using the service to “build a competing product or service, including to train competing AI models” or “reverse engineer or duplicate” the services. This change in OpenAI’s access to Claude comes as the ChatGPT-maker is reportedly preparing to release a new AI model, GPT-5, which is rumored to be better at coding.
OpenAI was plugging Claude into its own internal tools using special developer access (APIs), instead of using the regular chat interface, according to sources. This allowed the company to run tests to evaluate Claude’s capabilities in things like coding and creative writing against its own AI models, and check how Claude responded to safety-related prompts involving categories like CSAM, self-harm, and defamation, the sources say. The results help OpenAI compare its own models’ behavior under similar conditions and make adjustments as needed.
“It’s industry standard to evaluate other AI systems to benchmark progress and improve safety. While we respect Anthropic’s decision to cut off our API access, it’s disappointing considering our API remains available to them,” OpenAI’s chief communications officer Hannah Wong said in a statement to WIRED.
Nulty says that Anthropic will “continue to ensure OpenAI has API access for the purposes of benchmarking and safety evaluations as is standard practice across the industry.” The company did not respond to WIRED’s request for clarification on if and how OpenAI’s current Claude API restriction would impact this work.
Top tech companies yanking API access from competitors has been a tactic in the tech industry for years. Facebook did the same to Twitter-owned Vine (which led to allegations of anticompetitive behavior) and last month Salesforce restricted competitors from accessing certain data through the Slack API. This isn’t even a first for Anthropic. Last month, the company restricted the AI coding startup Windsurf’s direct access to its models after it was rumored OpenAI was set to acquire it. (That deal fell through).
Anthropic’s chief science officer Jared Kaplan spoke to TechCrunch at the time about revoking Windsurf’s access to Claude, saying “I think it would be odd for us to be selling Claude to OpenAI.”
A day before cutting off OpenAI’s access to the Claude API, Anthropic announced new rate limits on Claude Code, its AI-powered coding tool, citing explosive usage and, in some cases, violations of its terms of service.
Google Chrome is testing a new way to measure Core Web Vitals in Single Page Applications (SPAs), which is a long-standing blind spot in performance tracking that affects SEO audits and ranking signals.
Starting with Chrome 139, developers can opt into an origin trial for the Soft Navigations API. This enables measurement of metrics like LCP, CLS, and INP even when a page updates content without a full reload.
Why This Matters For SEO
SPAs are popular for speed and interactivity, but they’ve been notoriously difficult to monitor using tools like Lighthouse, field data in CrUX, or real user monitoring scripts.
That’s because SPAs often update the page using JavaScript without triggering a traditional navigation. As a result, Google’s measurement systems and most performance tools miss those updates when calculating Core Web Vitals.
This new API aims to close that gap, giving you a clearer picture of how your site performs in the real world, especially after a user clicks or navigates within an app-like interface.
What The New API Does
Chrome’s Soft Navigations API uses built-in heuristics to detect when a soft navigation happens. For example:
A user clicks a link
The page URL updates
The DOM visibly changes and triggers a paint
When these conditions are met, Chrome now treats it as a navigation event for performance measurement, even though no full page load occurred.
The API introduces new metrics and enhancements, including:
interaction-contentful-paint – lets you measure Largest Contentful Paint after a soft navigation
navigationId – added to performance entries so metrics can be tied to specific navigations (crucial when URLs change mid-interaction)
Extensions to layout shift, event timing, and INP to work across soft navigations
How To Try It
You can test this feature today in Chrome 139 using either:
Local testing: Enable chrome://flags/#soft-navigation-heuristics
Origin trial: Add a token to your site via meta tag or HTTP header to collect real user data
Chrome recommends enabling the “Advanced Paint Attribution” flag for the most complete data.
Things To Keep In Mind
Chrome’s Barry Pollard, who leads this initiative, emphasizes the API is still experimental:
“Wanna measure Core Web Vitals for for SPAs?
Well we’ve been working on the Soft Navigations API for that and we’re launching a new origin trial from Chrome 139.
Take it for a run on your app, and see if it correctly detects soft navigations on your application and let us know if it doesn’t!”
Here’s what else you should know:
Metrics may not be supported in older Chrome versions or other browsers
Your RUM provider may need to support navigationId and interaction-contentful-paintfor tracking
Some edge cases, like automatic redirects or replaceState() usage, may not register as navigations
Looking Ahead
This trial is a step toward making Core Web Vitals more accurate for modern JavaScript-heavy websites.
While the API isn’t yet integrated into Chrome’s public performance reports like CrUX, that could change if the trial proves successful.
If your site relies on React, Vue, Angular, or other SPA frameworks, now’s your chance to test how well Chrome’s new approach captures user experience.
Donald Trump has said that he has deployed nuclear-capable submarines to the “appropriate regions” in response to a threatening tweet by Russia’s former president Dmitry Medvedev, suggesting that he would be ready to launch a nuclear strike as tensions rise over the war in Ukraine.
In a post on Truth Social on Friday, Trump wrote that he had decided to reposition the nuclear submarines because of “highly provocative statements” by Medvedev, noting he was now the deputy chairman of Russia’s security council.
Medvedev had earlier said that Trump’s threats to sanction Russia and a recent ultimatum were “a threat and a step towards war”.
“I have ordered two Nuclear Submarines to be positioned in the appropriate regions, just in case these foolish and inflammatory statements are more than just that,” Trump responded. “Words are very important, and can often lead to unintended consequences, I hope this will not be one of those instances.”
He did not specify whether he was referring to nuclear-powered or nuclear-armed submarines.
In a video for the Guardian’s It’s Complicated, Josh Toussaint-Strauss investigates how the connection between civil nuclear power and nuclear weapons spans decades and continents as well as exposes siphoning of public money and the origins of the Iran nuclear program.
Medvedev, who was sidelined when Vladimir Putin returned to the presidency in 2012, is also an avid fan of X, formerly Twitter, where he often posts aggressive and curiously worded attacks against western countries and leaders in the evening in Moscow.
Earlier this week, Medvedev had attacked Trump for shortening his timeline for Russia to make progress toward peace with Ukraine from 50 days down to just 10, saying that he was ready to impose sanctions and other financial penalties against Russia if it didn’t comply.
“Trump’s playing the ultimatum game with Russia: 50 days or 10,” Medvedev wrote in a post. “He should remember 2 things: 1. Russia isn’t Israel or even Iran. 2. Each new ultimatum is a threat and a step towards war. Not between Russia and Ukraine, but with his own country.
“Don’t go down the Sleepy Joe road!” he added, referring to the former US president Joe Biden.
skip past newsletter promotion
after newsletter promotion
Trump has voiced frustration with Putin, who he said had been stalling on Trump’s efforts to broker a ceasefire between Russia and Ukraine, a campaign promise that he said that he could achieve in just 24 hours. On Thursday he described Russia’s continued attacks on civilian areas as “disgusting”.
“I go home. I tell the first lady, ‘You know, I spoke to Vladimir today. We had a wonderful conversation.’ She said, ‘Oh, really? Another city was just hit,’” he said in at the White House last month.
Putin has not responded to Trump’s ultimatum. On Friday, he said he wanted a “lasting and stable peace” in Ukraine but gave no indication that he was willing to make any concessions to achieve it, after a week in which Russian missiles and drones again caused death and destruction across Ukraine.
“We need a lasting and stable peace on solid foundations that would satisfy both Russia and Ukraine, and would ensure the security of both countries,” said Putin, speaking to journalists on Friday, a week before a new deadline imposed by Trump for hostilities to cease.
Putin has periodically claimed to be interested in peace, but only on terms wholly unacceptable to Kyiv. Last week, the third round of direct talks between Russia and Ukraine took place in Istanbul, but broke up in less than an hour and have so far led to no agreements except on prisoner exchanges.
In an apparent reference to Trump’s comments, Putin said on Friday: “As for any disappointments on the part of anyone, all disappointments arise from inflated expectations. This is a well-known general rule.”
After announcing the end of his long-running podcast, the show will go on for Marc Maron.
The comedian and Stick actor will release a new stand-up special, Marc Maron: Panicked, on HBO and HBO Max on Friday, Aug. 1, marking his second for the Warner Bros.-owned network. Cable subscribers can watch online on any live TV service that carries HBO, while cord-cutters can stream the show after it airs on cable HBO Max and other platforms, including on Prime Video and the Disney+, Hulu and HBO Max bundle.
At a glance: How to watch Marc Maron: Panicked comedy special online
Per the special’s logline, the comedian and podcaster “offers up his nuanced perspective on our increasingly uncertain world.” In a statement, Maron says, “I feel that this is the best work I’ve done. Everything came together. The direction, the production design, the shirt and the bits. HBO gets me and I’m thrilled to be presented by them.”
(HBO Max returned to its original name after a two-year stint as Max, while its parent company, formerly Warner Bros. Discovery, will split into two as Warner Bros. and Discovery Global.)
Watch the official trailer below and keep reading for more details on how to watch Marc Maron’s new comedy special with and without cable, plus the best HBO Max subscription deals.
How to Watch Marc Maron: Panicked Comedy Special with Cable Online
Marc Maron’s new comedy special, Marc Maron: Panicked, premieres on HBO on Friday, Aug. 1, at 8 p.m. PT/ET, and will be available to stream on HBO Max.
Cable subscribers can also watch the show online by logging into hbomax.com with their TV provider credentials, including DirecTV, Hulu + Live TV, Sling, Verizon, Xfinity and others.
Cord-cutters can watch the comedy special on any live TV streaming service that carries HBO or offers HBO Max as an add-on, including DirecTV, Hulu + Live TV and Sling.
How to Watch Marc Maron: Panicked Comedy Special Without Cable Online
Cord-cutters can stream Marc Maron: Panicked on Aug. 1 with a subscription to HBO Max, which costs $9.99 monthly with ads or $16.99 to $20.99 per month without ads.
You can add HBO Max to other streaming services, including Prime Video and DirecTV, or get it as part of a bundle with Hulu and Disney+ starting at $16.99.
Cord-cutters can watch the comedy special online through HBO Max, including on Prime Video, DirecTV and the Disney+, Hulu, HBO Max bundle. Keep reading for more details on all of the options for streaming HBO and HBO Max.
DirecTV
Watch Marc Maron: Panicked live on HBO on DirecTV, which offers a five-day free trial for eligible new subscribers. The live TV streaming service carries HBO and more than 90 other news, entertainment and sports channels starting at $59.99 for your first month for a limited time. Learn more about DirecTV’s packages here.
Eligible DirecTV subscribers can get two free months of HBO Max and stream HBO and Max original series (including The White Lotus, Hacks, The Last of Us and other Emmy-winning shows) and Warner Bros. films on demand. Learn more about the promo here.
Disney+, Hulu and Max Bundle
Watch Marc Maron: Panicked on demand on HBO Max and expand your library when you bundle the streamer with Disney+ and Hulu starting at $16.99 per month with ads — a savings of more than 40 percent monthly compared to subscribing to each service individually.
In addition to original HBO and HBO Max TV shows and Warner Bros. films, the bundle includes access to Disney, Pixar, Marvel and Lucasfilm (Star Wars and Indiana Jones franchises) content on Disney+; and Hulu and FX series such as The Handmaid’s Tale, The Bear and What We Do in the Shadows. Learn more about the Disney+, Hulu and HBO Max bundle here.
Hulu
New subscribers can watch Marc Maron’s new comedy special on HBO for free with a three-day trial to Hulu + Live TV. The live cable streaming service offers over 90 news, entertainment and sports channels as well as Disney+ and ESPN+ starting at $82.99 per month with ads; learn more here.
Sling
Stream Marc Maron: Panicked on demand on Sling’s Blue plan and pay only $30.50 for your first month when you add HBO Max — a savings of half off the first month of service (reg. $50.99 monthly) and a $5 off the HBO Max add-on every month (reg. $16.99 monthly). Sling Blue offers over 40 entertainment, sports and news networks; learn more about other packages here.
How to Watch HBO Max for Free: Best Subscription Deals, Trials
HBO Max currently does not offer a trial period, but there are a few ways to get the streaming service for free or at a discounted rate. DirecTV subscribers get two months of HBO Max for free, while DoorDash DashPass members get HBO Max with ads for free. Learn more about the best HBO Max streaming deals here.
