Anthropic has rewritten his 25,000-word “Constitution” for Claude. It is the manual for how AI should behave

Anthropic has published a completely renewed version of the so-called “Claude Constitution”. Yes friends, an AI also needs a constitution, or at least a series of documents that explain with total transparency what direction the company has decided to take with its AI tool. It is a way to save us trouble in the event that become aware. The document The question in question consists of 80 pages and nearly 25,000 words, and basically shows what values ​​Anthropic relies on to train its models and what they hope to achieve with it. Alluding to Asimov, it would be something like a broader and more complex version of his three laws of robotics. Why it is important. Anthropic carries a good time trying to differentiate from OpenAI, Google or xAI, wanting to position itself as the most ethical and safe alternative on the market. This Constitution is the centerpiece of their training method called “Constitutional AI”, where the model itself uses these principles to self-criticize and correct its responses during learning, instead of relying exclusively on human feedback. The document is not written for users or researchers: it is written for Claude. It was time to update. The first version of the Constitution, published in 2023, was a list of principles drawn from sources such as the UN Universal Declaration of Human Rights or, as they mention from Fortune, from Apple’s terms of service. Now, according to Anthropic, they have taken a completely different approach: “To be good actors in the world, AI models like Claude need to understand why we want them to behave in certain ways, rather than simply specifying what we want them to do,” affirms the company in its statement. The new document is structured around four fundamental values, and the most interesting thing is that Claude must prioritize them in this order when they conflict: Be largely secure: Do not undermine human AI oversight mechanisms during this critical phase of development. Be broadly ethical: act honestly, according to good values, avoiding inappropriate, dangerous or harmful actions. Comply with Anthropic guidelines– Follow specific company instructions when relevant. Be genuinely helpful: benefit the operators and users with whom it interacts. The majority of the document is concerned with developing these principles in more detail. In the utility section, Anthropic describe to Claude as “a brilliant friend who also possesses the knowledge of a doctor, lawyer and financial advisor.” But it also sets absolute limits, called “hard constraints,” that Claude must never cross: not provide significant assistance for bioweapon attacks, not create malware that can cause serious harm, not assist in attacks on critical infrastructure such as power grids or financial systems, and not help “kill or incapacitate the vast majority of humanity,” among others. Consciousness. The most striking part of the document appears in the section titled “The Nature of Claude,” where Anthropic openly acknowledges its uncertainty about whether Claude could have “some kind of conscience or moral status.” “We are concerned about Claude’s psychological safety, sense of identity, and well-being, both for Claude’s own sake and because these qualities may influence his integrity, judgment, and safety,” they count from the company. The company claims to have an internal team dedicated to “model well-being” that examines whether advanced systems could be sentient. Amanda Askell, the Anthropic philosopher who led the development of this new Constitution, explained told The Verge that the company doesn’t want to be “completely dismissive” about this issue, because “people wouldn’t take it seriously either if you just said ‘we’re not even open to this, we don’t investigate it, we don’t think about it.’” The document also raises complex moral dilemmas for Claude. For example, it states that “just as a human soldier might refuse to shoot peaceful protesters, or an employee might refuse to violate antitrust law, Claude should refuse to assist with actions that concentrate power in illegitimate ways. This is true even if the request comes from Anthropic itself.” And now what. Anthropic has published the entire Constitution under a Creative Commons CC0 1.0 license, meaning anyone can freely use it without asking permission. The company promises to maintain an updated version on its website, considering it to be a “living document and a continuous work in progress.” Cover image | Andrea De Santis and Anthropic In Xataka | Company CEOs say AI is saving them a day of work a week. Employees say otherwise

His kamikaze plan has rewritten the war manual

A year after the start of the Russian invasion of Ukraine, a drone instructor had an idea that sounded to science fiction– Pilot cheap quadcopters in order to ram and destroy other drones in mid-flight. Thus, what began as a joke between soldiers, “too much Star Wars”, they saidbecame in less than a year the spine of the Ukrainian defense. The origin. Given the shortage of anti-aircraft missiles and the russian waves of Iranian Shahed who put out cities, Ukrainian engineers and pilots they started redesigning commercial quadcopters to convert them into hit-to-kill interceptors. They were born out of necessity: Winter, power outages, and the inability of conventional defenses to process hundreds of low-cost threats pushed improvisation to become in doctrine. Crowdfunding programs like Come Back Alive and the initiative Dronefall They articulated production, training and logistics, financing and coordinating local manufacturers. How they work and their effectiveness. These interceptors require three conditions: speed and maneuverability to reach targets at hundreds of km/h, vision and guidance systems (from night cameras to semi-automatic guidance) and an explosive charge or kinetic capacity sufficient to destroy the threat upon impact. Models like the sting or variants by Wild Hornets They combine powerful propellers, thermal chambers and light warheads; The tactic is simple in concept, but extremely demanding in execution: detect, locate, launch and maneuver in windows of minutes before the attacker leaves range. Production and economy. lor we have told before, the strategic attractiveness it’s economical: an interceptor can cost between 2,500 and 6,000 dollarsin front of the million per missile of advanced systems. Multiple manufacturers, from Ukrainian SMEs to supported startups by Brave1allow scalability. Ukraine aims to produce hundreds and eventually thousands per dayIn fact, they are already reported thousands of interceptions and programs that connect twenty producers to standardize parts, training and supply. Field operations. Furthermore, the deployment requires a short chain: detection by radar or surveillance, link to a pilot or semi-autonomous system and launch with a very short margin of time (teams report 10-minute windows to intercept). Not only that. The effectiveness depends on the skill of the pilot (specialized courses show low pass rates) and the quality of the data link. When interceptors are not fully autonomous, the human variable remains the bottleneck: well-trained pilots achieve success rates much older. The Sting is much smaller than a typical Shahed drone Diversity of designs. Here the family of interceptors is heterogeneous: there are models that directly impact (ramming), designs with warhead projected at high speed, and guided drones optical sensor similar to small missiles. Plus: some are detachable and transportable in backpacks, and others are mass launchable from containers. This diversity allows the response to be adapted to the profile of the attacker (versus the slowness of a Shahed vs the speed of a Geran-3) and the operational environment. Results and effectiveness. Ukrainian reports speak of massive interceptions: hundreds killed in major attacks and aggregate figures of thousands of kills attributed to programs like Dronefall. Success rates vary (from 30% to 90% depending on the system, the class of the target and the expertise of the crew), but the economic impact is clear: replacing a defense missile with dozens or hundreds of cheap interceptors preserves strategic resources and forces Russia to inflate its operating costs. An interceptor crew prepares a Sting drone from their civilian vehicle Implications. NATO considers interceptors as a valuable complement to traditional layers of defense. The UK has already committed to co-producing interceptors for Ukraine; tests in allied airspace (e.g. trials in Denmark) demonstrate interest in integrating these solutions in territorial defense and protection of critical infrastructure. The main lesson for Europe is the need for cheap and scalable solutions to mass threats, not just high-cost, high-precision systems. Technical limitations. Not everything is optimism: interceptors also face scope problemsresistance to electronic interference and the ability to reach drones at very high altitudes or extreme speeds. The advent of reactor versions of the Shahed (Geran-3) that far exceed the speed of current interceptors forces the improvement race: greater propulsion, better sensor and autonomy, or alternatives such as higher-cost kinetic defense. Furthermore, dependence on human pilots with limited training conditions the sustainability of the effort. The next phase. Given the Russian advance towards faster drones, Ukraine and its partners are already working on new generations: faster interceptors, more robust sensors, semi-autonomous solutions and integrated deployments with radars and missiles depending on the objective. In parallel, non-kinetic defenses are being explored: from lasers to microwaves and EW systems that can complement or replace physical interceptors when speed or altitude exceed their capabilities. Strategic balance. If you will, the most profound change that interceptors introduce It’s doctrinal.: modern air warfare can be won by mass affordable and distributed response, and not just by expensive and one-off systems. Ukraine has shown in this sense that the combination of local manufacturing, civil financing and tactical adaptation transforms a weakness (lack of missiles, especially external) in operational advantage. The final caveat, however, is that this advantage it’s temporary: The adversary adapts, the technology scales, and the survival of the approach requires continued investment in design, production, and training. Image | Wild Hornets In Xataka | The crazy number of drones has turned the Ukrainian sky into the M-30 at rush hour. Identifying the enemy is a danger In Xataka | While Europe builds its Russian anti-drone wall, each nation loads its artillery: some with lasers, others with shotguns

We have just rewritten the genealogy of the inhabitants of the Cantabria of the Paleolithic. Thanks to DNA found in the mud

If we want to enter tens of thousands of years ago in the genetics of our ancestors, the only path we have is to study the bone remains in search of the low genetic material remaining in them. Or at least that used to be the case. 46,000 years ago. Because a new study has achieved Back 46,000 years in the past by analyzing sedimentary challenges in a Cantabrian cave. The genetic material found showed a common genetic ancestry among the inhabitants of the cave and populations settled in southern France in the same era. The mirror and his red lady. The history of the study of the archaeological site of the cave of El Mirón, located in Cantabria, begins with excavations initiated in 1996. Archaeological prospects would give one of its main fruits in 2010, with the discovery of the red lady of El Mirón. It was a partial skeleton that belonged to a woman from between 35 and 40 years deceased about 19,000 years ago. The appellation is due to the fact that the bone remains were found covered with a red ocher paint whose origin would not have been in the direct environment of the cave. Sedadna. One of the most striking details of the study is in its methodology. The new analysis of the remains of the site did not focus on the bone remains of the red lady, even on other types of bones. Instead they noticed the mud. The Sedadna methodology focuses on the sedimentary remains of DNA that still keep genetic information about the ancient inhabitants of the cave, humans or animals. These remains, Explain the study responsible for the study They allow us to ride ourselves very prior to that of the Red Lady, up to 46,000 years ago, in the Musteriense era, when the Neanderthals still inhabited Europe. From Fournol to Cantabria. However, the most relevant period in the new study is the Solutrense, the period in which the last maximum glacier occurred, approximately between 25,000 and 21,000 years ago. The sedadna extracted in the strata of this era allowed genetically to link the human populations that inhabited the cave in this era with other human groups. Specifically with the one known as Fournol lineagea group that we link with some deposits in Spain and in France. The new study allows us to better trace the genealogies of the human groups that inhabited the Peninsular North in Paleolithic. A genealogy that also covers the ancestry of the Villabruna lineage, which would have reached the region in bass Magdaleniensefrom the Balkans and northern Italy. The details of the study were published In an article In the magazine Nature Communications. Great carnivores. Another study key is in animals that inhabited the cave and, therefore, that they could be found in these times among the Iberian fauna. The team found traces belonging to carnivores such as leopard and hyenas, as well as the Dole, Wild Dog Asian or Indian Wild Dog (Cuon Alpinus), A canid now present in Southeast Asia. They also found DNA belonging to ungulates such as mammoths and rhinos, in addition to deer. In Xataka | “Look dad, ox”: the curious story of how an eight -year -old girl unwittingly discovered the paintings of the Altamira cave Image | University of New Mexico

Log In

Forgot password?

Forgot password?

Enter your account data and we will send you a link to reset your password.

Your password reset link appears to be invalid or expired.

Log in

Privacy Policy

Add to Collection

No Collections

Here you'll find all collections you've created before.