South Korea plans free nationwide AI access
South Korea’s science ministry chose three consortia led by SK Telecom, KT and Kakao to build chatbots and public service agents that every citizen can use free, with no usage limits.
The government gives the groups a combined 512 Nvidia B200 GPUs this year and will cover part of the running costs from 2027. Beta testing starts in September, full launch by December.
Rules require at least half of each service’s model usage to run on certified Korean models. So Seoul is buying American chips to lower its dependence on American AI.
OpenAI is breaking up with Cursor
OpenAI plans to remove its models from Cursor on November 12, 2026, following the coding assistant’s acquisition by SpaceXAI, though Cursor can choose to end access sooner.
OpenAI cited concerns about Elon Musk’s companies following its terms, pointing to Musk’s admission that xAI used OpenAI outputs to train models through a technique called distillation.
Cursor says OpenAI models serve just 5% of its customers and hopes to reach a deal, while Anthropic has already promised more compute for Claude inside the coding tool.
Meta tests robots to run its AI data centers
Meta is testing robots inside the data centers powering its AI, with machines handling tasks like swapping networking cables, rebooting servers, moving equipment racks, and checking hardware for problems.
Meta is testing machines from Watney Robotics, Kinova, and ABB, but humans still supervise them because tangled cables, limited batteries, obstacles, and visual inspections remain difficult.
One worker estimated a reliable cable-swapping robot could handle up to 80% of certain jobs, but that was not Meta’s forecast, and today’s robots still work slower than humans.
Anthropic says AI researchers now beat human ones
Anthropic published a paper Friday showing automated AI systems improved model performance on all 10 alignment benchmarks they were given, without degrading overall performance. Fellow Chen Yueh-Han led the work.
The systems copy how human researchers work. They search literature, propose a method, train the model for 30 minutes, then keep what works and discard what fails across repeated rounds.
Anthropic puts the automated researcher at roughly $4 an hour against $150 for its human ones. The paper notes the approach only works if the benchmarks measure the right things.
Google expands AI Overviews above search results
Google has started automatically expanding AI Overviews for some searches, placing full AI-generated answers and an “Ask anything” box above traditional links and pushing regular search results farther down the page.
Google says its systems expand summaries when they predict users will find them useful, though the behavior varies by query, browser, and session, making the rollout inconsistent for now.
Google says expanded answers encourage deeper follow-up searches and stop opening once users start scrolling, but publishers may care more about how much farther their links now sit below AI.
OpenAI weaponizes its API against SpaceX
Sam Altman just turned Elon Musk’s newest acquisition into collateral damage in tech’s most bitter personal feud.
SpaceX spent $60 billion buying the startup behind Cursor, only for OpenAI to trigger a change-of-control clause weeks later.
OpenAI openly blamed Musk’s history of breaking contracts as the explicit reason for ending the deal by November 12.
Cursor’s leadership is scrambling to negotiate behind the scenes after assuming OpenAI’s backend would remain neutral ground.
Musk sued OpenAI for $150 billion earlier this year claiming they abandoned their nonprofit roots, but a jury threw the case out because he waited too long to file.
Developers who built workflows around OpenAI’s models inside Cursor will soon get forced onto alternative backends.
If your team relies on Cursor, start benchmarking your prompts against Anthropic’s Claude or SpaceX’s native Grok models today so your dev pipeline does not stall out in November.
South Korea makes AI access a national right
South Korea is handing its entire population free, token-unlimited AI access to keep domestic tech giants from getting swallowed by American and Chinese platforms.
The government selected consortia led by SK Telecom, KT, and Kakao to roll out national AI agents for taxes, healthcare, and education.
Seoul is handing out 512 Nvidia B200 GPUs and bankrolling operating costs to guarantee unlimited access with zero subscription fees through 2028.
At least 50% of the underlying infrastructure must run on homegrown Korean models, creating a massive, state-backed anchor market for local tech.
Around 25% of South Koreans already pay for consumer AI tools compared to just 2% of Americans, making this massive rollout a natural evolution for a tech-heavy population.
Treating AI access as a fundamental civic right rather than a private subscription will instantly level the digital divide and force Western tech giants to reevaluate pricing.
If you build consumer AI products, keep a close watch on how South Korea integrates autonomous agents into government services like tax filings and medical bookings to see what nationwide AI adoption actually looks like at scale.
Nvidia moves beyond silicon to lock in AI infrastructure
Hyperscalers like Amazon and Google are spending billions building custom silicon, but Nvidia is quietly shifting its strategy to ensure competitor chips do not break its stranglehold on AI infrastructure.
Big tech labs are starting to manufacture their own inference chips, threatening Nvidia’s historic dominance as the sole provider of high-end silicon.
Nvidia is responding by buying up the software layer, pouring billions into open-weight model platforms like Hugging Face and Poolside to control where developers deploy their code.
The biggest technical bottleneck in modern AI data centers has shifted from pure processor compute to internal traffic flow, memory routing, and system orchestration.
Early in the AI boom, possessing state-of-the-art GPUs was the only metric that mattered for hardware dominance.
As rival chipmakers catch up on raw hardware, Nvidia is turning the whole data center into a tightly bound ecosystem that makes replacing its silicon nearly impossible.
If you are evaluating your team’s long-term cloud costs, audit how deeply your infrastructure depends on Nvidia’s proprietary software stacks versus open frameworks to avoid getting locked into a single ecosystem as custom silicon options mature.