Category: Top News

IBL News | New York

Beijing-based AI pioneer Zhipu AI released GLM-5.2 this week, a low-cost, powerful, open-source LLM featuring a massive 1-million-token context window that aims to attract users from US rivals dissatisfied with high prices and Washington’s abrupt order to suspend top American models overseas. Its stock rocketed in the Hong Kong market.

The launch underscored how Chinese players are seeking to capture users who are seeking alternatives to top models from Western leaders.

Zhipu’s announcement came shortly after San Francisco-based Anthropic abruptly suspended access to its flagship models Fable-5 and Mythos-5 to all foreign nationals, following a federal export-control directive issued on national-security grounds.

GLM-5.2 will be available to all users of Zhipu’s new GLM Coding Plan subscription, which is priced at just a tenth of Anthropic’s premium Claude Code and Claude Max tiers. The GLM-5.2 application programming interface (API) went live this week, while the model itself has been open-sourced under the permissive MIT license.

GLM-5.2 features native integrations with more than 20 popular agent tools, including developer mainstays Claude Code, Cline, and Cursor.

Introducing GLM-5.2: Frontier Intelligence, Open Weights

– Significant improvements in coding and agentic tasks
– Strong long-horizon capabilities with a 1M context window
– Two levels of reasoning effort: GLM-5.2 (max) pushes the limits, while GLM-5.2 (high) strikes a strong… pic.twitter.com/SjGPSVhePJ

— Z.ai (@Zai_org) June 16, 2026

To run the GLM-5.2, a 1.51TB AI model running on any laptop, Unsloth Studio just made it a lot less impossible, shrinking it by 84%. This way, it runs locally on a 256GB Mac

Unsloth fixed the quantization of the large GLM-5.2 model, dropping the 2-bit version, reducing the size to 238GB, and retaining ~82% accuracy.

To run it locally, users will need:

• • A 256GB Mac (M3/M4 Ultra) for single-box setup
  • Or a 24GB GPU plus 256GB RAM with memory offloading
  • Tools: llama.cpp, LM Studio, or Unsloth Studio

These offline coding agents and long-context tasks previously required a cloud API. Now, with Unsloth’s solution, no usage fees, no rate limits, and no data leaving anyone’s machine.

IBL News | New York

The nonprofit organizations ETS, Khan Academy, and TED are planning to launch the Khan TED Institute, which plans to offer degrees for under $10,000. The learning and career pathways will be developed in partnership with leading companies, including Google, Microsoft, Accenture, Bain & Company, McKinsey, and Replit.

The inaugural program will be organized around three pillars:

• Core knowledge in mathematics, statistics, economics, computer science, science, history, and writing.
• Applied AI skills, including AI‑assisted app development, financial modeling, building AI agents, and team‑based deployment projects.
• Communication and leadership,  developed through structured collaboration, peer tutoring, dialogue sessions, and public speaking.

In this partnership, Khan Academy participates with its learning platform, TED with its curated community of thinkers and transformative ideas, and ETS’s trusted expertise in measuring skills.

Applications are expected to open in 10 to 16 months.

“The Khan TED Institute is an extension of our vision to higher education, creating new paths that help more people find meaningful ways to contribute to the world around them,” said Sal Khan, founder and CEO of Khan Academy.

Amit Sevak, CEO of ETS, explained, “This collaboration helps open new pathways into the AI economy where skill‑based measurement becomes the critical link between learning and livelihood.”

Logan McClure Davda, CEO of TED: “At a time when learning is changing quickly, we now need new ways to help people engage with ideas, develop judgment, and apply what they learn. The Khan TED Institute is an effort to expand what education can look like in the age of AI.”

IBL News | New York

Spotify will introduce, in the coming weeks, as a research preview, a Google NotebookLM – style stand-alone desktop app, named “Studio by Spotify Labs,”  designed to explore a topic and, by following personal context, create a podcast about it.

Announced during its 2026 Investor Day, this podcast-generation tool has an agent that can browse the web and fetch personal information to create a personal and private podcast.

The tool can create a daily briefing or a podcast based on the user’s email, calendar, and documents.

All these AI-generated podcasts are saved in the user’s Spotify Library for personal consumption and are not available publicly.

Since Google’s NotebookLM was released, the format for creating podcasts to explore a topic or get daily briefings has been adopted by companies like Adobe and ElevenLabs, as well as apps like Hero and Huxe.

Also, Claude Code or Codex has debuted a command-line tool for users of coding tools to create personal podcasts and save them to their Spotify library.

The company said, “Studio by Spotify Labs understands your Spotify taste across music, podcasts, and audiobooks, while also drawing on world knowledge to help you find the audio you want faster. With your permission, it can take action on your behalf: researching topics, using a web browser, organizing information, and helping complete tasks. It can also connect to the tools you use every day, like your calendar, inbox, and notes.”

IBL News | New York

OpenAI announced it reached the first national-level cooperation with any Government in the world. It happened with Malta, a small European country with a population of 550,000.

Starting in May, Maltese citizens who complete an AI literacy course will obtain one year of free usage of ChatGPT Plus (with a monthly fee of $20),

This course, developed by the University of Malta, covers the basic principles of AI, along with methods for responsible use in family and work scenarios.

ChatGPT currently has more than 900 million weekly active users.

George Osborne, head of OpenAI for Countries, said, “Intelligence is becoming a national public utility. Malta has taken the lead, and we hope other countries will follow.”

The Maltese experiment might determine whether “the state pays for all citizens’ AI use” is a replicable public service innovation or an expensive marketing campaign.

Trying to make AI a daily, basic tool, OpenAI’s strategy seems, in this case, to be based on acquiring future customers in batches first through government channels.

The company is currently promoting cooperation in the education system in Estonia and Greece, and these countries are the most likely to follow in the first wave.

IBL News | New York

Anthropic said on Friday that the Federal Government had ordered the company to suspend access to its Fable 5 and Mythos 5 AI models for all foreign nationals.

The company revealed the order in a social media post, saying that it “must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance.”

Foreign nationals — including citizens of allied nations like Canada or Britain — who work for Anthropic on it are also affected by the suspension.

The order to limit access to the new AI models came from the Commerce Department.

These limits on foreigners represent a new chapter in Anthropic’s often rocky relationship with the federal government.

It is also the second time the Trump administration has targeted the San Francisco-based company.

Government restrictions on the sale of key technologies, such as computer chips, to countries like China are common.

But the order limiting who can use Anthropic’s technology is unusually expansive and could, in theory, block Anthropic employees who are.

IBL News | New York

Anthropic revealed that 80%+ of the code merged into its codebase as of May 2026 is authored by Claude. 3 (in February 2025, this number was in the low single digits), while it detailed its progress toward recursive self-improvement and its implications.

“At Anthropic, we are delegating a growing share of AI development to AI systems themselves, which is speeding up our work.”

The company believes that the so-called recursive self-improvement practice – with AI systems capable of fully autonomously designing and developing their own successors— could arrive sooner than most institutions are prepared for.

Today, autonomous agents can now run code themselves and delegate hours of work to other agents.

Systems capable of fully building their own successors would represent a major development in science, healthcare, and beyond, but might increase the risk of humans losing control over AI systems.

In the future, agents could become capable of building and training models themselves. If this happens, future versions of Claude could be continuously improved by Claude itself.

The Anthropic Institute is showing that AI is already accelerating the development of AI systems. To take just one example: today, Anthropic engineers on average ship 8x as much code per quarter as they did from 2021 to 2025.

The technical trends discussed in this piece suggest that AI systems are going to become much more capable in the coming years. These trends have huge implications.

The evidence points out that the rate at which AI models improve is accelerating, with the length of tasks doubling roughly every four months, up from an earlier trend of doubling every seven months.

In March 2024, Claude Opus 3 could complete software tasks that would take humans about 4 minutes. A year later, Claude Sonnet 3.7 managed tasks that would have taken about an hour and a half. A year after that, Claude Opus 4.6 managed 12-hour tasks.1 If this trend holds, tasks that take a skilled person days could come within range this year. In 2027, AI systems could be capable of tasks that take a person weeks.

Our internal data shows Claude is accelerating AI development—a possible path to recursive self-improvement, or AI autonomously building a more capable successor.

It’s happening faster than we thought, and the implications deserve greater attention. https://t.co/OVVPJO7VQx

— Anthropic (@AnthropicAI) June 4, 2026

IBL News | New York

The Chinese model MiniMax M3, natively multimodal and with a context window of up to 1M tokens, was released this month. It’s an open-weight model, meaning its trained parameters are publicly available for download, local deployment, and fine-tuning.
However, the company withheld the original training code, training data pipelines, and specific inference operators. It can run on a desktop computer.

Users can use the model via MiniMax M3, MiniMax Code, the Token Plan, and API services.

According to the company,

• “On SWE-Bench Pro, which measures coding capability, MiniMax M3 surpasses GPT-5.5 and Gemini 3.1 Pro and approaches Opus 4.7. On SVG-Bench, a benchmark that comprehensively evaluates SVG generation performance, MiniMax M3 surpasses Opus 4.7.”“On OmniDocBench, a multimodal benchmark, MiniMax M3 scores above Gemini 3.1 Pro. On Claw-Eval, an end-to-end evaluation framework for autonomous agents, MiniMax M3 achieves the highest score.”

• “For long-horizon complex tasks, MiniMax Code’s Agent Team can break large tasks down into multi-stage, concurrent, and dynamically adjustable workflows, which are then advanced collaboratively by a cluster of agents. Through a Producer + Verifier adversarial harness loop, the Agent Team can continuously produce, reflect, and correct itself during execution. It can run autonomously for days without human intervention and ultimately deliver high-quality results.”

• “We have seen that Claude Code has also recently released Dynamic Workflows in a similar direction. Compared with Claude Code’s stronger emphasis on fixed orchestration based on JS code, MiniMax Code focuses more on “deep reflection and continuous error correction”: the agent adjusts its plans and priorities in real time based on task progress, while users can step in at any time to add requirements or correct the direction.”

On the other hand, the company announced that MiniMax Code, “built on a harness based on the outstanding open-source community projects OpenCode and Pi”, will be open-sourced this project in the future.

IBL News | New York

Paris-based AI company Mistral AI is positioning itself as a critical European counterweight to U.S. dominance in the race for superintelligence or AGI. The start-up is growing by playing to its home crowd: selling services hosted in European data centers, independent of American and Chinese competitors, to local governments and enterprises.

Experts say that U.S. tariffs and President Trump’s threats to take over Greenland have given new impetus to a European push for tech independence.

Guillaume Lample, Mistral’s co-founder and chief scientist, said, “Very soon in the future, we are probably going to see AGI or superintelligence, so it is very important that we have access to these models also in Europe.”

“Also, as long as we have adversaries that are threatening, we do need to have our own capabilities,” he said. “Europe in particular needs to have strategic autonomy when it comes to defense systems.”

Founded by three French AI researchers from Google and Meta Platforms, Mistral was valued at around $14 billion last year and promises more than $1 billion in revenue this year, a fraction of the size of its Silicon Valley rivals.

Mistral has already landed deals to supply AI services to European corporate heavyweights such as Airbus and BMW, as well as the French military. Mistral also said it was building a new 10-megawatt data center south of Paris, part of the $4.7 billion it is plowing into data centers in France and Sweden.

However, the company acknowledges that a big obstacle to tech independence is the scale of investment necessary.

“We can’t put 50 billion [dollars] on the table to build a gigawatt ahead of demand,” the executive said. “That’s potentially our biggest bottleneck.”