AI Daily Roundup: Palantir's Karp Torches AI Labs, AWS Backs Vibe Coding, Apple Ships Siri AI, Design Arena Raises $7.9M, and Congress Bets Big on ChatGPT
Palantir CEO Alex Karp accuses AI labs of "Marxist" tactics during a record quarter, AWS strikes a multi-year deal with vibe-coding startup Superblocks, Apple finally launches an AI-powered Siri that feels both impressive and overdue, Design Arena raises $7.9M to crowdsource "taste" for AI models, and US Congress funnels 90% of its AI spending to OpenAI's ChatGPT.
1. Palantir CEO Alex Karp Calls AI Labs "Marxist" — While Posting Record-Breaking Results
In his quarterly shareholder letter, Palantir CEO Alex Karp delivered a blistering critique of AI frontier labs, accusing them of trying to "capture the means of production of their purported partners." Karp — who holds a PhD in social theory — drew explicit Marxist parallels, arguing that enterprises building on closed AI lab APIs risk letting those labs migrate their intellectual property, expertise, and institutional knowledge into proprietary models that ultimately compete against them.
The irony: Palantir itself just had its best quarter ever. The company reported $1.9 billion in revenue (up 93% year-over-year) and $1.1 billion in profit — "more profit in a single quarter than we did in total revenue in the same period the year before," Karp wrote. Palantir positions itself as model-agnostic enterprise software, letting organizations control their own data and AI orchestration. Karp's underlying point — that AI labs may use customer data and access to build competing products — echoes similar warnings from Microsoft CEO Satya Nadella in recent weeks. The data is becoming clear: as the AI agent market matures, enterprises are increasingly wary of vendor lock-in at the model layer.
2. AWS Partners With Vibe-Coding Startup Superblocks — Enterprise AI Apps Inside Private Clouds
Amazon Web Services has announced a multi-year joint marketing agreement with Superblocks, a vibe-coding startup that lets business users build internal AI applications without sending data to external model providers. Under the deal, Superblocks will be embedded within AWS customers' private clouds — apps spin up Amazon Aurora databases internally, integrate with AWS's Bedrock AI gateway, and automatically fall under IT's existing security and audit controls.
Superblocks co-founder Brad Menezes told TechCrunch that enterprise demand for open-source and multi-model setups has flipped dramatically: "60 days ago they were like, 'I want a specific model. It's called Anthropic.' Now they want open models and flexibility." Open-weight models already account for 29% of traffic routed through popular multi-model management tools. AWS will also help sell Superblocks through its Marketplace, while the startup — which has 50 employees and $60 million in total funding — gains instant credibility with enterprise buyers. The partnership symbolizes a broader trend: hyperscalers want enterprises to buy AI scaffolding on their clouds while using whichever models they choose, keeping the cloud provider at the center of the stack.
3. Apple Finally Ships an AI-Powered Siri — But It Feels Like a Long-Overdue Fix, Not a Revolution
The AI-powered Siri has arrived in the consumer beta of iOS 27, and it actually works. Users can now have natural back-and-forth conversations, ask Siri to find personal context buried in their photos and emails, launch apps and websites by voice, brainstorm ideas, and even identify real-world objects through the camera viewfinder. Apple used its partnership with Google to train proprietary foundation models that run on Apple silicon and its private cloud compute infrastructure.
Yet as TechCrunch notes, the arrival feels "anticlimactic." While Siri is now a genuinely functional AI assistant, the industry has already moved far beyond chatbots — AI agents are completing multi-step tasks, coding software, and working alongside humans using their computers. A competent voice assistant feels more like Apple finally fixing a long-standing bug than doing something revolutionary. The real test will come when Siri launches to the general public alongside iOS 27 in September, and whether Apple can keep pace with the agentic AI capabilities that competitors are already shipping.
4. Design Arena Raises $7.9M Seed Round — Crowdsourcing "Taste" as the Missing AI Benchmark
Intelligence, the company behind Design Arena — an AI evaluation platform used by 5.3 million people worldwide — has announced a $7.9 million seed round led by Index Ventures with participation from Conviction, A*, and Valkyrie. The platform lets users rank and compare AI-generated visual content (websites, images, games) side by side, providing frontier labs with scalable human feedback on design quality — what co-founder Grace Li calls "the missing bottleneck for models to make improvements in the design space."
The business is already generating $60 million in annual recurring revenue, making it one of the fastest-growing AI infrastructure companies in the evaluation space. Users log in to compare outputs, and the platform tracks how aesthetic preferences differ across regions and evolve over time — noting, for example, that web dashboards in Asia tend toward a more maximalist design style. The raise comes as the broader human-evaluation market heats up: LM Arena raised $150 million in its Series A in January, while earlier competitor Yupp shut down despite having 1.3 million users. For AI agent builders, platforms like Design Arena provide a crucial signal that automated benchmarks alone cannot capture.
5. US Congress Spends 90% of Its AI Budget on ChatGPT — A Signal for Enterprise Adoption
According to a CNBC analysis of House disbursement records, OpenAI's ChatGPT received roughly 90% of all spending on AI tools by House offices, committees, and institutional accounts during the year ending March 31. Congress spent approximately $100,580 across 798 transactions on ChatGPT, out of at least $113,740 in total AI tool spending. Democratic offices accounted for $54,165 — three times the $15,782 spent by Republican offices.
Congressional staffers are using ChatGPT and other AI tools for drafting memos, summarizing and analyzing legislation, responding to constituents, preparing hearing materials, sorting through policy research, and drafting social media posts. While the raw dollar amounts are modest, the 90% concentration is a striking data point. It suggests that even in the most security-conscious and reputation-sensitive environments, a single platform has captured near-total mindshare. For enterprise AI adoption, the signal is clear: the platforms that win the "first use" battle tend to dominate, and the network effects of institutional adoption are powerful.
📰 TechCrunch · CNBC