In partnership with
News of the day
1. OpenAI previews its GPT‑5.6 series (Sol, Terra, Luna), with major gains in coding, biology and cybersecurity, plus a hardened safety stack. → Read more
2. Anthropic's AI advancements mean junior engineers are no longer needed, potentially causing an economic shock as other industries adopt similar tech. → Read more
3. EPFL's new AI model mimics the human brain, offering greater user control and transparency beyond traditional 'black box' LLMs. → Read more
4. Amazon drops OpenAI movie, data center workers protest, and Meta pauses employee tracking after data leak. AI's film industry impact and Anthropic's government relations are also discussed. → Read more
Our take
Hi Dotikers!
A few days ago, GPT-5.6 was being framed as a tariff weapon, with a passing note that Anthropic's equivalent models had been pulled from the market by export controls. The asymmetry seemed to play in OpenAI's favor. What came next arrived faster than expected: today, it's OpenAI that's under watch.
The company is launching the GPT-5.6 series in preview. Three models: Sol at the top, Terra in the middle, Luna at entry level. But the real story isn't the lineup, it's how it's shipping. At the request of the US administration, OpenAI is opening access only to a handful of hand-picked partners, whose list is passed along to the government. According to a memo from Sam Altman, Washington will sign off on access client by client during the test phase. A wider rollout would follow a few weeks later, if all goes well.
The reason comes down to one word: cyber. OpenAI is touting a model able to spot vulnerabilities and the basic building blocks of an exploit, on par with Anthropic's Mythos Preview while using three times fewer tokens. Except the Mythos models have been frozen since June 12, unplugged on orders from that same government. Measuring yourself against a product nobody can use makes for a comfortable benchmark. As for the rest, OpenAI is signing the arrangement while repeating that this kind of control shouldn't become the norm.
It's the first time Washington has blocked the launch of an American model before the fact. The executive order signed this month was meant to be a voluntary step; in practice, it's now a case-by-case approval, and the framework that's supposed to govern all of it is being written live, over the phone with the Secretary of Commerce. OpenAI named its lineup after the sun, the earth, and the moon. One very earthly detail remains: you now need the state's blessing to sell a tool that knows how to rummage through your neighbor's browser.
Alex.
HR and IT need to work as one. Here's how
Every missed onboarding step, delayed offboard, or broken provisioning handoff has a root cause: HR and IT aren't aligned. This guide gives both teams a shared framework for the full employee lifecycle.
Meme of the day
