RT @deepseek_ai
🚀 DeepSeek-V4 Preview is officially live & open-sourced! Welcome to the era of cost-effective 1M context length.

🔹 DeepSeek-V4-Pro: 1.6T total / 49B active params. Performance rivaling the world's top closed-source models.
🔹 DeepSeek-V4-Flash: 284B total / 13B active params. Your fast, efficient, and economical choice.

Try it now at http://chat.deepseek.com via Expert Mode / Instant Mode. API is updated & available today!

📄 Tech Report: https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro/blob/main/DeepSeek_V4.pdf
🤗 Open Weights: https://huggingface.co/collections/deepseek-ai/deepseek-v4

,"Citrini’s argument is that AI enables the automation of some (eventually most) services, recreating Baumol’s cost disease within the service sector. Services that can be broken down into discreet tasks (“Taylorized”) and can make use of an increasingly data rich environment, such as call center work, basic accounting, legal discovery, graphic design, much sales work, or routine diagnostics and coding, will be automated, reducing the total labor force employed and increasing productivity. At the same time, there will remain a labor-intensive service subsector with low productivity growth. This labor-intensive subsector will itself be under immense pressure as AI and robotics advance.

This bifurcation recreates Baumol’s cost disease within the service sector, destroying many of the well-paid positions that have retained some degree of workplace autonomy in the process. The result of this transformation in work would be the emergence of an economy shaped by very few highly paid service workers, and an army of low-skill, low-paid workers. All of this would take place against the backdrop of a collapse in the total mass of service employment due to productivity gains.

The lesson from technological innovation in the manufacturing sector is that increased productivity means that firms require fewer workers. While new markets may develop along with new services, these new services will not escape the division between a small number of well-paid workers and a dwindling mass of their low-paid peers. The worst-case scenario would be one in which even this low-waged work disappears thanks to service automation."

https://jacobin.com/2026/04/ai-stagnation-services-productivity-unemployment

#AI #GenerativeAI #SecularStagnation #Automation #Productivity #Unemployment

"Immigration and Customs Enforcement (ICE) just inked a $12.2 million contract for an artificial intelligence tool that claims to map out immigrants’ daily routines, habits, and real-time location and categorize them as potential threats, per procurement records reviewed by the Lever.

Dubbed “Project SAFE HAVEN,” the product is advertised by defense vendor Edge Ops LLC as a “question-based AI interface” that uses “persistent passive data collection” to map “patterns of life,” a surveillance tactic that ICE says will identify the “habitual locations, routes, and behavioral patterns” of its targets.

Additional features of the technology described in procurement documents include real-time location tracking and analysis that will categorize individuals and groups as affiliated with ostensible criminal organizations, such as gangs or cartels. That includes building “target profiles” that track individuals’ activity by linking data obtained from Wi-Fi network connections and mobile smart devices, such as cell phones and smartwatches.

A promotional blurb for the tool on Edge Ops LLC’s website claims that Project SAFE HAVEN “transforms the way we identify, locate, and map illegal migrants.”"

https://jacobin.com/2026/04/ice-contract-ai-surveillance-immigrants/

#USA #Trump #Authoritarianism #ICE #AI #Surveillance #Privacy

"The US federal government’s central energy information agency is planning to implement a mandatory nationwide survey of data centers focused on their energy use, according to a letter seen by WIRED. This survey would be the first effort of its type to collect basic information about data centers.

The letter was sent to senators Elizabeth Warren and Josh Hawley on April 9 by the head of the Energy Information Administration, Tristan Abbey, and comes in response to a previous inquiry from the senators about the EIA’s plans to get more information about data centers. WIRED reported on Hawley and Warren’s letter last month.

"Americans deserve to know how much energy data centers are sucking up and what that's doing to their utility bills,” Warren told WIRED in a statement. “The EIA's mandatory survey is an important first step towards holding data centers accountable, but people are hurting right now. I'm pushing EIA to collect and share this data as soon as possible."

The EIA told WIRED that it doesn’t have any specifics to share beyond what is in the letter to the senators."

https://www.wired.com/story/the-us-government-to-ask-data-centers-how-much-power-they-use/

#USA #DataCenters #Energy #BigTech

"In January, Hanaa’ Tameez and I broke the story that The New York Times, The Guardian, and USA Today Co. had begun limiting the Wayback Machine’s access to their news articles. Our reporting showed that these decisions, including a “hard block” by the Times that started late last year, were driven by publishers’ concern that the Internet Archive’s free library of webpage snapshots could be scraped by AI companies to train their commercial models.

Now, journalists and digital rights nonprofit organizations are pushing back against this trend and advocating for news publishers to lift their restrictions.

On Monday, Wired first reported on the publication of a new petition organized by the digital rights nonprofit Fight for the Future. The open letter does not call for any specific policy from publishers, but “applauds” the Wayback Machine for its work “at a time where many major media outlets are questioning whether to allow the Wayback Machine to continue to preserve journalism.” The petition has already been signed by over 120 journalists, including Cory Doctorow, Taylor Lorenz, and Ron Suskind.

“The Internet Archive is a national treasure. I use it daily, and have for many, many years. I cannot imagine doing the work I do without it,” MS Now host Rachel Maddow wrote in a testimonial published alongside the letter."

https://www.niemanlab.org/2026/04/journalists-champion-wayback-machine-after-news-publishers-limit-article-archiving/

#News #Journalism #Media #DigitalRights #InternetArchive

RT @AkshatRathi
Half of US data centers planned for 2026 are expected to be delayed or canceled. One big reason is shortage of electrical equipment, such as transformers, switchgear and batteries. US doesn't have manufacturing capacity, forcing it to rely on imports. 🎁🔗

https://www.bloomberg.com/news/features/2026-04-01/us-ai-data-center-expansion-relies-on-chinese-electrical-equipment-imports

"The point is: let the agent do the boring stuff, the stuff that won't teach you anything new, or try out different things you'd otherwise not have time for. Then you evaluate what it came up with, take the ideas that are actually reasonable and correct, and finalize the implementation. Yes, sure, you can also use an agent for that final step.

And I would like to suggest that slowing the fuck down is the way to go. Give yourself time to think about what you're actually building and why. Give yourself an opportunity to say, fuck no, we don't need this. Set yourself limits on how much code you let the clanker generate per day, in line with your ability to actually review the code.

Anything that defines the gestalt of your system, that is architecture, API, and so on, write it by hand. Maybe use tab completion for some nostalgic feels. Or do some pair programming with your agent. Be in the code. Because the simple act of having to write the thing or seeing it being built up step by step introduces friction that allows you to better understand what you want to build and how the system "feels". This is where your experience and taste come in, something the current SOTA models simply cannot yet replace. And slowing the fuck down and suffering some friction is what allows you to learn and grow.

The end result will be systems and codebases that continue to be maintainable, at least as maintainable as our old systems before agents. Yes, those were not perfect either. Your users will thank you, as your product now sparks joy instead of slop. You'll build fewer features, but the right ones. Learning to say no is a feature in itself."

https://mariozechner.at/posts/2026-03-25-thoughts-on-slowing-the-fuck-down/

#AI #LLMs #AIAgents #AgenticAI #SoftwareDevelopment #Programming

"Imagine that automating a task requires a $10 million dollar investment (buying the software, onboarding, connecting it to the rest of the system, etc.). In one case, this task is the only non-automated task left in a job; in the other case, if this task is automated, there are 19 other non-automated tasks left. The firm has a much higher incentive to automate the task in the first case than the second because it can then replace the worker and reap the cost savings involved.1

Because of this, firms have a stronger incentive to invest in technology to automate low dimensional jobs. In a low-dimensional job, automating all or most of the core tasks can eliminate the position and the wage bill altogether. That makes the return to automation much larger. In other words, not all “unexposed” tasks matter equally: in some jobs the remaining tasks still keep the existing worker at the firm; in others they do not.

This gives a clear prediction: even if a job is not currently “exposed” to AI, in the sense that AI is not being used for the tasks involved, if it is low dimensional and the technology is getting close to automating the tasks, it should be considered at risk. Firms will work harder and invest more to automate the task(s) involved than in the case where jobs have many non-automated tasks.

This is why we think people should be more worried about jobs like trucking and warehousing."

https://aleximas.substack.com/p/how-will-ai-driven-automation-actually

#AI #GenerativeAI #Automation #Unemployment #Productivity

RT @TheHackersNews
Telegram blocked 43M+ channels in 2025, yet threat actors stayed.

Yochai Corem shows they adapted—rebuilding in days, gating access, and shifting sensitive comms off-platform while keeping Telegram for scale.

🔗 How criminals evolved despite Telegram’s crackdown → https://thehackernews.com/expert-insights/2026/03/telegrams-crackdown-changed-how-threat.html

"Many of today’s popular new media forms—podcasts, Substacks, and social media feeds—rely for their techniques and their content on the old medium of long fact. The popular podcast Freakonomics Radio was derived from the host’s 2005 book. The pioneering Serial podcast used the techniques of narrative nonfiction writing in audio form—and emerged out of public radio’s This American Life, whose founder, Ira Glass, edited the 2007 anthology The New Kings of Nonfiction, which linked audio storytelling with literary storytelling of prior eras. Reams of streaming documentaries and fictional dramas that strive for verisimilitude (such as Succession and The Diplomat) are made by showrunners and screenwriters versed in narrative nonfiction. Shows from Morning Joe to Rachel Maddow present authors as experts alongside policymakers and elected officials, and Maddow is herself the author of four narrative nonfiction books. Opening the broadcast, she often relates an episode drawn from the history books—literally—and then pointedly joins it to the present.

Those figures generally present books that are recognizably “on topic.” That’s good and necessary. But through my own work as an author and teacher, I’ve been struck by the pertinence of nonfiction books that don’t deal directly with current affairs. These books develop narratives that at first glance are well outside the news cycle, but as you read them, you find they speak powerfully to the moment precisely because they don’t succumb to the presentism, partisanship, and winners-and-losers schemas too often regarded as inviolable norms of media today."

https://newrepublic.com/article/207659/non-fiction-publishing-threat-important-ever

#Books #BookPublishing #NonFiction #NonFictionBooks

"Cuban Vice Prime Minister Óscar Pérez-Oliva Fraga announced Monday a sweeping package of measures expanding the economic role of Cubans residing abroad that would allow them to own and invest in private business for the first time since the revolution, including the possibility of large-scale investments by foreign capital in key sectors of the economy, as well as the participation of Cubans residing abroad in the national financial system.

“We are not talking about specific businesses. The doors of our country are open to the participation of the Cuban community residing abroad,” Pérez told the Cuban news program Mesa Redonda in an interview.

Fernandez do Cossio said in his interview with Drop Site, that the scope of what is now being offered far exceeds what has been possible in the past. While it’s hard to imagine Costco or Starbucks on the island, he said that the freedom of political movement that would come from a lifting of the embargo would open up new possibilities in terms of economic and political reforms. Trump himself long ago registered the trademark for Trump Hotel Havana, according to Cuban government records, and a new agreement would pave the way for such a development.

Pérez, who also serves as minister of foreign trade and investment, said in his comments that Cuba was taking steps to reduce red tape for foreign investment and added that Cubans abroad would also be afforded the opportunity to invest in development projects and funds in Cuba.

The official said Cuba would welcome investment by U.S. firms but specified that it is U.S. law that does not permit this."

https://www.dropsitenews.com/p/cuba-prepared-offer-lump-sum-agreement-united-states-property-lost-revolution?triedRedirect=true

#Cuba #USA #Trump #Imperialism #NeoColonialism