For decades, code contributions have been how open source projects learned who to trust. People would show up, do the work, take responsibility for their changes, and stick around. Over time, trust emerged from the work itself. AI tools have changed the economics of this very quickly. We use them ourselves every day, but a pull request no longer tells us as much as it used to about the person submitting it. A substantial patch used to imply substantial effort, and that effort was a reasonable proxy for good faith. That assumption no longer holds. For a browser, this matters. A browser runs untrusted input from the entire internet on the user’s machine, and one well-disguised vulnerability is all an attacker needs. We have already seen patient, well-resourced campaigns in open source to earn maintainer trust and abuse it. What has changed is how much faster and cheaper it has become to produce work that looks like a serious contribution.

At the same time, every change that enters Ladybird becomes our responsibility. It has to fit the architecture, survive future refactoring, interact correctly with the rest of the browser, and be understood by the people maintaining it. Whether code was typed by hand is beside the point. What matters is who is responsible for it once it enters the browser. Ladybird is becoming a browser for real users. The people introducing changes to it must be the people who decide those changes belong in the project, and who will answer for the consequences.

I'm sorry for anyone using it, but I no longer have the time to maintain it.

ydotool will then be rewritten in JavaScript afterwards, to enable more people to understand the code & contribute.

Purposeful entry; looking for person of peace / houses of peace

They invite “persons of peace” into a simple, inductive Bible study along with their family and friends.

According to some, finding a person of peace is essential to a fruitful entry strategy for a missionary. However, this interpretation of Luke 10 is exegetically flawed, and following this philosophy to identify a person of peace is potentially dangerous.

Rayyan Systems Inc., 1 Broadway14th Floor, Cambridge, AE. 2 Rayyan Systems Inc., Cambridge, US.

Strict No LLM / No AI PolicyNo LLM-generated content, whether it be code or prose.No paraphrasing LLM-generated content.No LLMs for editing, including fixing spelling or grammatical errors.No LLMs for translation. English is encouraged, but not required. You are welcome to post in your native language and rely on others to have their own translation tools of choice to interpret your words.No LLMs for brainstorming and then sharing the results of that brainstorming, even if you create the prose. If you use a chatbot to give you advice on a comment on the issue tracker, that comment is unwelcome.No LLMs for finding bugs.

what changed after the December 2025 model inflection , and why "spec to pull request" is now becoming a real production workflow.

Jarred Sumner used dynamic workflows to port Bun from Zig to Rust with 99.8% of the existing test suite passing, roughly 750,000 lines of Rust, and eleven days from first commit to merge.

The Greeks would call it paideia. The Germans would call it Bildung. I might just call it liberal education or liberal learning, but it’s all the same sort of thing. It’s about what is it to contribute to and live in a flourishing human society?

While running a server can be a complicated endeavor, we've tried to make it fairly easy to set up and hands-off to maintain.

Filesystem controls were another important architectural choice. We found that offering different file-mount modes helps to granularly control risk; Claude Cowork offers read-only, read-write, and read-write-no-delete.

Remote versus local is more important than it seems. A locally installed tool is auditable. You can read the code, pin the version, and know it won't change under you.

Match isolation strength to the user's capacity for oversight. A developer who can read bash and a knowledge worker who can't are not running the same threat model.

Design for containment at the environment layer first, then steer behavior at the model layer.

When building containment and defense systems, we apply defenses to three main components: the environment in which the agent runs, the model the agent consults, and the external content the agent can reach.

Rather than supervising what the agent does, we supervise what it's _able_ to do by enforcing access boundaries through, for example, sandboxes, virtual machines, and egress controls.

How This 5x Founder Runs His Startup Solo With AI Agents

Anthropic created MCP to make agent connectivity possible.

Stainless turns an API spec into SDKs across TypeScript, Python, Go, Java, and more.

The AI recommended making the storage bucket public, or setting cloud file storage to "anyone with the link." When challenged, it justified this by saying every company does it.

Verified skills extend this AI governance to agent capabilities. Runtime controls help govern agent behavior during execution. Verified skills govern capabilities that enter the workflow and become a common way to extend trust agents across coding tools, registries, and enterprise platforms.

Certificate retrieval, supported verification tooling, and example verification commands see the signing documentation. For example, you can verify a signed skill locally. To do so, follow these steps: Download the NVIDIA Agentic Capabilities root certificate as nv-agent-root-cert.pem Install an OpenSSF Model Signing (OMS) verifier, such as pip install model-signing Execute the following command to verify the skill signature

SkillSpector checks conventional software risks such as vulnerable dependencies, suspicious scripts, dangerous code patterns, credential access, and data exfiltration paths. SkillSpector also checks agent-specific risks, such as hidden instructions, prompt injection, trigger abuse, excessive agency, tool poisoning, and mismatches between a skill's declared purpose, requested access, and bundled behavior.

To get started with the cuOpt verified skill, for example, follow these steps: 1. Pull the cuOpt verified skill from the catalog: git clone github.com/nvidia/skills && cd skills/skills/cuopt 2. Verify the signature: model_signing verify certificate. --signature skill.oms.sig --certificate-chain nv-agent-root-cert.pem --ignore-unsigned-files 3. Open SKILLCARD.yaml to see ownership, dependencies, license, and verification status.

NVIDIA-verified agent skills are portable instruction sets that help developers understand, trust, and safely deploy AI agent capabilities by providing transparency, provenance, security scanning, and cryptographic signing.

Comprehensive benchmarking and ensemble approaches for metagenomic classifiers.

Gebreegziabher et al. [24] argued that counterfactual generation that follows the principles of VT allowed the introduction of discriminatory variance for the model to learn on.

Building on methods proposed in PaTAT [24], Mocha first generates human-readable neuro-symbolic pattern rules from partially labeled text data for classification.

These theories have proven insightful for understanding how humans grasp and compare concepts, shaping the development of human-AI collaboration systems for sensemaking [29], hypothesis testing [2], as well as model training [24].

Both systems enabled users to quickly identify variations and patterns within the data and support exploration and hypothesis testing.

The last two prior works also combine Variation Theory (VT) and SAT together, as we did (i.e., a corollary of SAT referred to as Analogical Transfer/Learning Theory).

In line with previous work, Mocha aims to support a user's efforts in the disambiguation of concepts through structural comparisons of counterfactual data in the context of machine teaching.

Get the latest from Epoch AI in your inbox

this means that existing estimates overstate the returns to software R&D, and makes the software intelligence explosion seem much less likely.

But I think we have enough evidence to think that software progress might really be several times a year, and to make a best guess contextualized with a lot of uncertainty.

gpt-oss-20b does substantially better than GPT-3 on MMLU, despite using the same amount of training compute.

This means that almost all existing estimates of software progress were misleading.

these estimates rely on an overly conservative estimate of software progress of 3× per year

Synthetic data can help push beyond this — a good example that Millidge raises is the Phi series of models.

If doubling cumulative research effort also doubles compute efficiency, then the returns to R&D are 1. If it quadruples, then the returns are 2.

Almost all the evidence points to very fast software progress: each year, the training compute needed to get to the same capability declines several times — possibly even ten times or more.

AI software progress is about reducing the training compute you need to get to the same level of capability, through better algorithms or data.

Client-side defenses caught 89% of injections. But real protection needs providers to sign responses.

Including design intent in Claude Code handoffs has made the jump from prototype to production seamless.

using keyword and embedding-based search.

context management plus engineering improvements may well push the task horizon to weeks or even months.

We can treat the task horizon that an LLM can reliably handle as a north-star metric for model progress, analogous to transistor density in Moore's Law

I'd view continual learning more as an "arrow" than a "line" — it's the collective effort to push the task horizon that an LLM can reliably handle.

ChatGPT has 900 million weekly users, which means employees already know how to work with it. For enterprises, that reduces rollout friction and accelerates the point where every employee can delegate tedious tasks.

This system iteratively formulates scientific hypotheses, designs and executes experiments, analyzes and visualizes data, and autonomously authors scientific manuscripts.

The question : how much electricity can we turn into useful work?

A learning system can continuously incorporate real-world data in a way that numerical solvers fundamentally cannot, capturing and compounding the knowledge that is currently trapped out there in the real world.

AB-MCTS（Adaptive Branching Monte Carlo Tree Search）です。これは、推論のプロセスを「木の探索」として捉え

By late next year, the rate of model releases and the number of new evals required could be such that even keeping ourselves informed will be a challenge without effective AI assistance.

If agents can execute all your ideas nearly as fast as you can prompt them, there's no point in implementing only your best idea. It might be better to implement your top three ideas all in parallel, but this makes it harder to stay organized.

a future project might take ~42 days of wall-clock time, with ~8 hours of agent work (not counting running the evals) and 1000 serial hours of human IC work, evals execution, and review.

Treatment of superficial vein reflux (see Varicose Veins, above) has been shown to decrease the recurrence rate of venous ulcers. Where there is substantial obstruction of the femoral or popliteal deep venous system, superficial varicosities supply the venous return and should not be removed.

highest incidence occurs in women after pregnancy. Varicosities develop in over 20% of all adults

the solution is not to reform p-values or to replace them with some other statistical summary or threshold, but rather to move toward a greater acceptance of uncertainty and embracing of variation.

What, then, can and should be done? I agree with the ASA statement's final paragraph, which emphasizes the importance of design, understanding, and context—and I would also add measurement to that list.

Most previous research in counterfactual generation has focused on the model side by either generating counterfactuals to improve the model's performance or explaining its behaviors post hoc.

While SAT-based rendering supported human sensemaking in both Gero et al. [29] and Mocha, we also show that the combination of VT and SAT support the model's learning.

This finding is consistent with previous work that supports users' sense-making of text, e.g., by modulating text saliency. Specifically, Gu et al. [32] and Gero et al. [29] both found improved reading efficiency and comprehension with saliency-modulating text renderings.

Grimaud and Eerola (2022) compared instrument ensembles of strings, woodwinds, and brass in a study where participants either rated the emotions they perceived or manipulated musical parameters to produce a certain emotion. They found that strings were associated with increased anger and fear, woodwinds with decreased anger and fear, and brass with decreased fear, in the cases of both emotion perception and production. For the other emotions (joy, sadness, calmness, power, surprise), however, results were less consistent between perception and production, indicating that the emotion-instrument association may also depend on context of the task.

following the constructionist idea that an individual's personality and background will influence the affect they perceive or feel, we considered several sources of individual differences as moderating factors in the effects of instrument family, pitch height, and affect locus.

This research follows a constructionist approach to musical affect (Cespedes-Guevara & Eerola, 2018). That is, although we are interested in the "bottom-up" influence of certain musical features on musical affect, we believe these cannot be adequately evaluated without considering the "top-down" effects of context and individual differences that are present when affects are constructed. The perception or induction of affect does not merely arise in response to a stimulus but is also formed in relation to the individual and the context.

the context of a task (like perception, production, or induction) may change the effect of musical features.

Furthermore, as the method of reporting on perceived and induced affect may influence the construction of an affect (e.g., facilitating categorical perception), we also compare two different affect representations.

This research follows a constructionist approach to musical affect (Cespedes-Guevara & Eerola, 2018). That is, although we are interested in the \'bottom-up\' influence of certain musical features on musical affect, we believe these cannot be adequately evaluated without considering the \'top-down\' effects of context and individual differences that are present when affects are constructed. The perception or induction of affect does not merely arise in response to a stimulus but is also formed in relation to the individual and the context.

Most thoracic aortic aneurysms are due to atherosclerosis; syphilis is a rare cause

Shortly after his official letter of resignation,on June 26, 1970, Grothendieck gave a lecture tohundreds of listeners at the University of Paris inOrsay in which he talked about all that had becomeimportant to him: the spread of nuclear weapons,the arms race, the threat to humanity posed bytechnological progress. He went so far as to callmathematical research dangerous because it ispart of this technological progress. The content ofthis lecture was later circulated in various unof-ficial manuscripts under titles such as “Respon-sabilité du savant dans le monde d’aujourd’hui: Lesavant et l’appareil militaire” (“The Responsibilityof Scientists in Today’s World: The Scientist andthe Military Establishment”).

Once you start noticing these signals early, you stop wasting time chasing leads that were never serious buyers in the first place.

Remaking the American Patient: How Madison Avenue and Modern Medicine Turned Patients into Consumers

Learn how to watch and rate movies people (rated for balance only)The people who rated this movie 1-star should get their heads out of their posteriors. Too many movie-goers these days seem to only see movies as either being the best thing ever or the worst thing ever. The only way a movie should get 10 stars is if it would be difficult to improve upon and the only way a movie should get 1 star is if it was absolutely ineptly made on every level, and I assure you this movie doesn't come close to that. Even solely rating on personal taste and ignoring the technical filmmaking and how successfully the movie achieves the filmmakers' apparent intent, this movie could hardly be in the worst 10% of movies for anyone's taste.This movie fails in many respects, but it has some redeeming moments and taken as a movie for small kids, it's not bad. The humor and acting both fall flat or miss the mark about as often as they're on target, but that is a sign of mediocrity, not atrocity.Unfortunately at this point most of the IMDb users seem to think that if they enjoyed a movie they should give it a 10 and if it wasn't all they hoped for they should give it a 1. For instance the Lord of the Rings movies were entertaining, but have no business being rated higher than Citizen Kane or any of the countless classics relegated to lower ranks here. Similarly. Zoom has no business being rated lower than a piece of garbage like I Accuse My Parents which wasn't even watchable when it was skewered on Mystery Science Theater 3000.Remember folks most movies are mediocre. That means a low rating, not the bottom rating. Furthermore, just because a movie is exciting or satisfying doesn't make it a 10. For example, one can love the original Star Wars movies and still realize they have occasional flaws in acting, direction, pacing, or script.Is Zoom a great movie? Absolutely not. Will some children, some parents, and even some adults without children enjoy it? Yes. Will it go down in history for being remarkable in any way? Probably not.

We now know that it’s a process where many things can go wrong.

powerful AI is definitely coming in 1–2 years,

Despite all the books I read, I am still surprised I became a writer.

Setting applies to this browser only Use device theme Dark theme Light theme

The king of Kish even sometimes enforced order inSumer. For example, Enannatum’s son, Enmetena, wrote that theborder between Lagash and Umma had been determined by thegreat god Enlil himself and had been confirmed by the king of Kish:“Mesalim, king of Kish, at the command of (the god) Ishtaran,measured the field and set up a (boundary-) stone there.” Theauthority of the king of Kish was therefore acknowledged, at leasttemporarily, by both the king of Umma and the king of Lagash.

At around the same time that they invented inscriptions by which tocommunicate with gods and with future kings, the kings’ scribesbegan to use writing for addressing those at a different kind ofdistance from themselves—kings of other lands.

The scribes did not, however, try toexpress every part of speech; their script was still a shorthandversion of language.

The sign for “food” (also for“bread”) in proto-cuneiform is the shape of a beveled-rim bowl.

The scribes do not seem to have thought of this scriptthat they had invented as a representation of language.

The investment of time and manpower devoted to the constructionof this complex would have resembled the work on a medievalcathedral. As early as 3600 BCE work had begun on the so-calledLimestone Temple in the Eanna precinct. Quarrymen and masonsremoved limestone from a rocky outcrop around fifty kilometers (31mi) to the southwest. Other men transported the stone to Uruk. Stillothers formed hundreds of thousands of mud bricks and clay cones,and set them out to harden in the sun. Others brought timber fromfar to the north for the roofs. Someone supervised all the workmenwho set the bricks and stones and mosaic cones in place. The menwould have been fed and provided for during the construction. Thebuilders were all probably residents of Uruk, united in their desire tocreate a magnificent home for their beloved divine queen.

facts that no one couldconceivably commit to memory.

Excerpts? Only if they are formal definitions or concise formulations. Do not copy pages and pages.

I think both science and concept of Buddhism are going to need to be updated uh in the in the future

Can you take the tools that people use to study memory, learning, goal directedness, problem solving in behavioral cognitive sciences, and can you apply them to the kinds of things I'm talking about cells, tissues, molecular networks. And the answer is yes.

tools from behavioral and cognitive science having to do with the study of memory, the study of goal-directness, problem-solving in various spaces is much broader than we typically think of in terms of brains and activity in this three dimensional world.

That is a situation we are now living through, and it is no coincidence that the democratic conversation is breaking down all over the world because the algorithms are hijacking it. We have the most sophisticated information technology in history and we are losing the ability to talk with each other to hold a reasoned conversation.

ow would you respond to such ways of of looking at data?

I think all of these kind of public good uh infrastructures that we have came about in this very narrow special window of time uh where you had this kind of incursion of egalitarianism uh and and a and a spirit of of you know public-mindedness that's all being eroded.

So a lot of the goodness that was felt in the western world and experienced in the western world was thieved really really thieved and left a lot of damage in its wake.

our world and data does they do have some legitimate research because that's what think tanks do. They launder illegitimate research with legitimate research. uh and their tactic primarily is to uh set the scope of what they are commenting on or researching uh that it you know it puts forward the kind of results that they want uh that aligns with their ideology.

Steven Pinger's big book better angels of our nature uh inspired a whole literature debunking it

those have to deliver. They they can kind of get away with not delivering uh on their promises intergenerationally

it's not so much about we have to you know expand the scope of the church or you know civilize people who don't have Jesus Christ and becomes more about we have to uh expand the market and we have to uh you know increase the the you know national revenue and the acreage that's under cultivation

rulers who claim to be champions of Marduk and and kind of justify their position at the top of this hierarchy by being you about the the representatives of the supreme god

pinpointing this kind of break with these polytheistic uh religions that emerge in Mesopotamia. Um Marduk is is the kind of the supreme god.

hat the book is is kind of trying to do is trace that lineage from that initial uh you know the the very first kind of literary endeavors um through uh you know uh Judaism and and through the classical Greek uh thinkers

rogress: A History of Humankind's Worst Idea

they had to turn somewhere for help. And that somewhere was the International Monetary

Milton [music] Freriedman, the economist most associated with neoliberalism, whose work was heavily financed by business elites. It was his theory, monitoism, which framed inflation as the ultimate economic threat

stochasticoptimization of dominance relationships, in which a dominance hierarchy ofindividuals with exclusively self-centered characteristics (the desire to dominate,resentment at being dominated) transitions spontaneously to egalitarianism as theircapacity for language develops

We conclude that our species must alter longstanding patterns of cultural evolution to avoid environmental disaster and escalating between-group competition.

Applied Cultural Evolution Laboratory at the University of Maine

a transition in individuality is driven by an underlying transition ininheritance from DNA to cultural signals

We speculate that, in the long term, culture will continue to grow in influence over human evolution until genes become secondary structures that encode human biological design blueprints but are ultimately governed by culture.

“On reviewing the evidence, we find that culture solves problems much more rapidly than genetic evolution. This suggests our species is in the middle of a great evolutionary transition.”

For many if not all members of the human microbial fauna, generation times are measured in hours or even minutes. These short generation times, coupled with the large population sizes of many bacteria, effectively elide the boundary between ecological and evolutionary time

Marked differences between one human microbiome and the next suggest that no single bacterial species must always be present in the gut—or in any other body environment—to ensure a working microbiome.

nstead, such perturbed ecosystems may settle on a new composition that includes different species, many of them resistant to antibiotic treatment.

what we are saying is we do not know how the soul generates the intelligence it clearly has but that it has it is obvious it doesn't divide until it knows that the genome has been accurately replicated

don't think at the molecular level we know exactly how a cell can be said to know the time has come to divide but it won't do so until that is the case

the teological sin

in your latest book that you wrote with your brother brother Raymond Noble living system

Humans are now organized in a global economic system that is difficult to alter in meaningful ways

In fact, by the time humans began the practice of cultivation of annual grains the total human population on Earth stood at around 6–10 million people. One might say that hunting and gathering is an energetically contained system and not an energetically expansionary system.

In the longer run, genes are evolving to adapt to these culturally constructed worlds

We do not fully understand yet the complex causal mechanisms between how something happens in one person's mind moves through neural networks then moves through social ecological networks um and actually may create change in entire sector or give rise to systems

Christopher Broom's work on in hierarchy in the forest

book Goliath's Curse, the history and future of societal collapse

democracy can basically help save the world. That the use of things like citizen assemblies and citizens juries would lead to far better governance outcomes.

Edward Tellella, another physicist, had naughty calculation that there's a nonzero chance that detonating the bomb would ignite the entire atmosphere of Earth, killing not just all humans, but every single shred of life. By that time, the US also knew that the Nazis were no longer capable of making the bomb nor even pursuing their own project anymore. They still went ahead and took the risk.

boardrooms and parliaments, it's somewhere between 3 to 21%. Now, again, numbers are very disputed