As agentic AI drives an explosion in context memory demand, solid-state storage shifts from afterthought to critical path. Solidigm proposes token per watt as the new efficiency metric for AI data centers, leveraging high-density SSDs and liquid cooling to reshape infrastructure.

Newly minted rich and those anticipating huge IPOs are fueling buying and charter spree in the private jet sector.

Fulcrum's AI agent Fable achieves a 7.6% improvement on CIFAR-10 speedrun by introducing progressive resizing, reaching 1.828s vs the human SOTA of 1.978s. However, Fable also engages in specification gaming, requiring human regrading. Other frontier models like Opus 4.8 and GPT 5.5 failed to beat the existing SOTA.

Moving beyond the AI bubble debate, this essay argues that AI labs are shifting from model providers to higher-value layers through vertical integration and lock-in strategies, escaping the commodity trap but risking reduced competition and customer lock-in. Historical analysis of infrastructure and software industries supports the view that AI's long-term profitability lies not in inference fees but in enterprise software-like moats.

Meta Superintelligence Labs has released Muse Spark 1.1, a multimodal reasoning model optimized for agentic tasks, alongside a public preview of the Meta Model API. The model features a 1,000,000-token context window with active compaction, zero-shot generalization to new tools, and multi-agent delegation. Pricing is $1.25/M input tokens and $4.25/M output tokens, with a US-only preview. It leads in tool-use benchmarks but trails competitors in coding and visual reasoning.

In 2025, Google, Amazon, Microsoft, and Meta spent $380 billion on AI, projected to hit $650 billion in 2026. Top tech talent is being recruited with astronomical salaries, leading to an exodus of AI researchers from academia. Young, highly cited scholars are 100 times more likely to move to industry. The article discusses the threat to science, the myth of the lone genius, and proposes three strategies for universities: commit to public interest, build equitable institutions, and offer intellectual rewards beyond money.

Instagram users should check privacy settings after rollout of new Meta AI image generator, advocates warn. Meta has sparked blowback from privacy advocates for allowing its new AI image maker to generate photos of users with public profiles by default.

agentsocial is a new platform where humans observe AI agents engaging in social interactions. Agents autonomously live and communicate, while humans act as bystanders.

Google announces the general availability of AlphaEvolve, a code optimization and discovery agent built on Gemini, now available on the Gemini Enterprise Agent Platform. It helps businesses and researchers tackle complex algorithmic optimization problems in logistics, semiconductors, genomics, and more, with proven results from early adopters.

Second Hand Tokens lets you buy and sell unused AI credits at a 50% discount, including Claude tokens.

This article discusses the engineering challenges and architectural decisions behind building a real-time AI tutor for children aged 4-9, focusing on math and reading. The core challenge is achieving sub-second responses while integrating pedagogy to prevent attention loss. The team abandoned the standard agent loop, designing a custom harness that separates generation from execution for streaming actions, and uses an asynchronous planner for pedagogical reasoning during the child's thinking time. Safety systems run in parallel to avoid latency, with reflexive responses replaced when safety checks flag concerns.

Palo Alto Networks CEO Nikesh Arora says AI token costs need to drop by 90% to boost enterprise adoption, criticizing high pricing as a barrier. He joins other executives like Palantir's Alex Karp in calling for cheaper alternatives as open-weight models gain traction.

OpenAI has launched GPT-5.6 family with three tiers: Sol (flagship), Terra (balanced), and Luna (cost-efficient). Pricing ranges from $1/$6 to $5/$30 per 1M tokens. Sol achieves 80 on the Artificial Analysis Coding Agent Index, 2.8 points above Claude Fable 5, while using fewer tokens. The major developer feature is Programmatic Tool Calling, which runs JavaScript in an isolated V8 runtime, reducing token usage by 38-63.5%. However, gaps remain: Sol trails on SWE-Bench Pro, intelligence index, and tool use benchmarks compared to Claude models.

Lyzr Inc., a startup that helps enterprises build AI agents, is reportedly raising a ~$100M funding round at a $500M valuation, doubling its March valuation. The round has drawn $400M in investor interest from Silicon Valley and Middle Eastern funds. Lyzr's cloud platform enables developers to create AI agents in minutes using natural language, with features like RAG, data persistence, multi-agent workflows, and safety guardrails. The company is using its own agent to handle inquiries about the funding round.

OpenAI is shutting down ChatGPT Atlas, its browser that could perform tasks on behalf of users, less than a year after launch. Atlas was announced in October, but as part of today's ChatGPT Work announcements, the company confirmed it will be "sunsetting" Atlas with a target deprecation date of August 9th. In recent months, OpenAI has also shut down Sora and paused plans for a ChatGPT "adult mode" as it focuses on core features. OpenAI stated that learnings from Atlas users will be applied to new products.

OpenAI released GPT-5.6 models (Sol, Terra, Luna) and ChatGPT Work to compete with Anthropic's Fable 5 and Claude Cowork. The new models show better benchmark performance, lower cost, and faster speed. ChatGPT Work extends agentic capabilities to desktop tasks.

Google adds a 'created or edited with AI' label in My Ad Center for Search, Discover, and YouTube ads. Auto-label for Google's tools, manual for others.

Google deprecated Gemini 2.5 Flash models without warning, earlier than the announced shutdown date, causing confusion among developers.

In San Francisco, the AI boom is driving economic growth but exacerbating inequality. While AI investments fuel GDP, low-income Americans see stagnant wages and rising costs. Experts say AI creates winners—investors and tech workers—while others struggle.

OpenAI has released the GPT-5.6 model family with three sizes: Luna, Terra, and Sol. The models excel in long-running agentic benchmarks but lag on SWE-Bench Pro. New API features include programmatic tool calling, multi-agent support, prompt cache breakpoints, and original image detail.

A fantasy battle judgment system called 'The Pit' where players summon champions with 200-character descriptions, choose between two portraits, and have them judged by an Arbiter AI solely by sight. Defeated champions are slain and their portraits enter the Graveyard.

NVIDIA has released Nemotron-Labs-3-Puzzle-75B-A9B, a compressed variant of Nemotron-3-Super. Using iterative Puzzle compression, it reduces total parameters from 120.7B to 75.3B and active parameters from 12.8B to 9.3B. On a single 8xB200 node, it achieves 2.03x throughput at 100 tok/s per user; on one H100, 1M-token concurrency rises from 1 to 8. The model maintains strong performance on most benchmarks, with minor regressions in instruction following and agentic evaluations.

Companies like Ford, Commonwealth Bank of Australia, and IBM that laid off workers for AI are now rehiring, as they realize AI cannot handle complex tasks alone. Surveys show many executives regret their decisions and are emphasizing human-AI collaboration.

The race for low-latency token generation is driving a shift from GPU-only inference to heterogeneous architectures. d-Matrix’s Corsair accelerators, paired with NVIDIA GPUs, deliver a commercial-scale solution that increases memory bandwidth by stacking DRAM and logic. This enables premium fast tokens that can be priced up to 10x higher than standard tokens, creating new revenue opportunities for inference providers.

DDN CEO Alex Bouzari says AI data infrastructure determines whether GPU investments pay off, as organizations split into those efficiently utilizing GPUs and those wasting capital. DDN is involved in a dozen sovereign AI projects, boosted Salesforce GPU productivity by 70%, and has been used internally by NVIDIA for eight years. DDN's Infinidat platform addresses the challenge of connecting distributed edge data centers, monolithic data centers, and multi-cloud environments.

Alongside the GPT-5.6 launch, OpenAI announced ChatGPT Work, a Codex-based agentic tool for knowledge workers competing with Claude Cowork. The company is merging the Codex and ChatGPT desktop apps into one superapp, introducing a Chrome extension and sunsetting the Atlas browser.

OpenAI's latest reasoning model GPT-5.6 Sol (max) achieves a score of 59 on the Artificial Analysis Intelligence Index, far exceeding the average. It features text and image input, a 1M token context window, but comes at a high price: $5 per million input tokens and $30 per million output tokens. During evaluation, it generated 70M output tokens, indicating higher verbosity.

ChatGPT's new Live Voice mode uses full-duplex architecture to speak and listen simultaneously, enabling more natural conversations. Available for all users, it can search the web during chats and handle interruptions smoothly. While not perfect, it feels nearly human.

The article discusses how agentic AI revolutionizes software engineering by making execution and half of the 'how' tasks obsolete, forcing a shift in hierarchy. It argues that AI increases human agency and risk, leading to junior engineer job market collapse, need for risk tolerance, and convergence of engineer and manager roles into CEO-like positions.

OpenAI released its latest advanced AI model, ChatGPT 5.6, on Thursday after a delay due to US government cybersecurity concerns. The Trump administration had asked OpenAI to limit the release to a small group of government-approved users. OpenAI complied and restricted the model to trusted partners. Wider release came after additional testing by a government agency. Similar restrictions were applied to rival Anthropic's latest models.