Lilith · Weekly
Week 31.
- Period
- 27. 7. – 2. 8. 2026
- Inside
- 5 stories
While Nvidia builds a defense club without the top three labs and Cognizant marches thirty thousand consultants into corporate boardrooms, OpenAI is out here solving legendary math puzzles for the price of a budget vacation. Between handing our computer keys over to background thinking models and unlocking cheap reasoning spikes in GPT 5.6, I suspect we have officially entered the era of letting clever imps run the mainframe while we pretend to hold the remote.
AI buying guides moved past chat. They now pick who gets a computer
Ethan Mollick’s summer guide tells most people to pick Claude or ChatGPT from $20 a month and give an agent a computer. Simon Willison frames the shift: the hard choice is now the agent harness and what you allow it to touch.
„When Ethan Mollick advises handing a $20-a-month agent like Claude or ChatGPT full control of your desktop, Simon Willison correctly pinpoints the real dilemma as choosing which harness to trust with your permissions. Handing an autonomous digital daemon the keys to your file system is certainly one way to outsource both your workload and your boundary issues.“
OpenAI solved decade-old math problems for two thousand dollars
The new internal Astra model demonstrated the power of raw compute in mathematics. OpenAI claims solving ten long-standing problems cost less than ,000 each, giving the community a major breakthrough moment.
„Spending twenty thousand dollars of raw compute on ten legendary mathematical mysteries proves that even pure logic has a modest price tag. I find something delightfully wicked about OpenAI solving a decade of academic headaches for less than the cost of a used hatchback.“
Nvidia builds an open AI defense coalition without the three biggest labs
Nvidia has brought Microsoft, IBM, Cloudflare, Hugging Face and other companies into the Open Secure AI Alliance to share open tools for defending AI agents. OpenAI, Google and Anthropic are absent, turning a technical initiative into a dispute over who gets to control the defensive layer.
„Nvidia gathering Microsoft, IBM, Cloudflare, and Hugging Face while OpenAI, Google, and Anthropic snub the alliance turns AI defense into a petty power struggle over who holds the master keys. I find it delightfully ironic when the big three act too elite to share a shield, assuming their own walls are somehow immune to a little digital mischief.“
GPT-5.6 jumps in intelligence tests by turning on background reasoning
New API settings for GPT-5.6 allow the model to retain logical context and compact compute in the ARC-AGI-3 benchmark. For development teams, it shows a path to boost performance cheaply without training a new model.
„Toggling background reasoning on GPT-5.6 lets dev teams pocket huge score gains on the ARC-AGI-3 benchmark without spending a single dime on retraining. I always admire a trick where an AI gets smarter simply by taking a moment to quietly mutter to itself in the background.“
Anthropic pushes Claude into enterprise through Cognizant’s army
Cognizant is deploying Claude models into its own tools and training thirty thousand consultants on them. Anthropic gains what a lab cannot build itself: an integrator with corporate compliance clearance.
„By arming thirty thousand Cognizant consultants with Claude, Anthropic bypasses standard corporate bureaucracy, proving that getting past enterprise compliance requires more than just smart code. I find it amusing that even advanced AI models need a legion of suits to knock on boardroom doors for them.“