Full-text search across article titles and summaries. Refine with sector, year, content type, and reading-time filters.
Sponsored
Search has changed, but most enterprises aren't keeping up. This Algolia whitepaper [5] is a technical deep-dive into the architecture that allows LLMs to reason about user intent, find relevant data, and execute using the right tools. Topics include: * The intricacies behind an indexed tool catalog * The ins and outs of ranking candidates under policy constraints
* How Algolia's retrieval and ranking layer powers enterprise intent routers * How Algolia turns natural language into safe, auditable action
OpenAI is rolling out Ultrafast today for GPT-6.1 Sol in the API, Codex, and ChatGPT Work. The feature offers near-Astra-level intelligence at up to 8x faster speeds than Sol Standard. It is built for work where speed and intelligence make a difference. API pricing for GPT-6.1 Sol in Ultrafast mode is $12 per million input tokens and $60 per million output tokens.
Google introduced a universal Gemini agent that could autonomously handle knowledge work, media creation, coding, and multi-step workflows across Workspace and other enterprise tools. It also added persistent context, multi-agent orchestration, model routing, governance, and cost controls.
Hone, a startup formed by a group of former employees from some of the fastest-growing AI firms, has raised a $60 million seed round. The startup aims to create AI agents that can help run a business, essentially serving as professional staffers. These agents will be able to field long-running tasks that run over the course of weeks or even months. The five-month-old startup joins a growing number of companies selling AI software to businesses that promise to automate complex work.
Epoch Automation Reports show that models can't yet autonomously produce Epoch-quality work. Frontier models performed reliably on well-defined tasks, but they failed in the more open-ended aspects that prevent full automation. Open-weight models lag further behind, struggling even on the well-defined tasks that frontier models handle reliably. The models struggle to pick up Epoch's standards, even with ample reference material, and while they could identify promising research directions, they lacked the judgment to successfully follow through.
Speculative Decoding is a technique where a small draft model proposes a likely draft sequence, and a big model verifies that sequence in parallel in a single forward pass. It enables the ability to look at many tokens at once and accept the ones that would have been generated by the big model. When the batch size is smaller than the optimal number of tokens, speculative decoding allows models to use 'free' compute to complete sequences faster. However, at high throughput, you could actually be wasting compute when you could be using that time to move memory.
TLDR is hiring its first PM to help build the agent-first operating layer used across the company. We're looking for a builder who has shipped real products/systems with LLMs. Click here to learn more [13].
Quicksand is an async Python API for launching, controlling, and snapshotting QEMU virtual machines with a particular focus on sandboxing AI agents. It provides pre-built Linux VMs for Ubuntu and Alpine distros. Quicksand supports x86_64 and ARM64 across macOS, Linux, and Windows. The sandboxes do not need root privileges or Docker.
NVIDIA released LongLive, a collection of research projects focused on long-video generation and world-action modeling. Each project includes its own code, documentation, and model weights.
ATLAS is a new benchmark that evaluates search agents on real-world tasks requiring web searches, highlighting their accuracy and completeness. The benchmark reveals that even high-effort agents miss significant golden answers, indicating major room for improvement in search engines. ATLAS differs from existing benchmarks by focusing on non-memorized tasks, requiring extensive multi-domain searches, and offering a cost-effective grading process.
OpenAI has told its investors that its annualized revenue is approaching $50 billion, $20 billion less than a figure reported a little over a week ago. The previously reported figure was devised by OpenAI investors as an attempt to produce a direct comparison with Anthropic's annualized revenues. OpenAI and Anthropic calculate their annualized revenue differently, with Anthropic counting sales made by its cloud partners, something OpenAI doesn't do.
Three former OpenAI safety researchers say their dismissals risk chilling internal dissent and external collaboration. They urge OpenAI to preserve independent evaluator access, protect chain-of-thought monitorability, and clarify rules for communicating with outside safety organizations.
Sponsored
In a flood of "AI-washing," it's hard to tell what's trustworthy. For enterprises with sensitive IP, there's Redwood RangerAI. Benefit from AI and agentic orchestration without the governance headaches. See how
Midjourney is now testing a new 'Thinking Mode' for image generation on its Alpha website.
Voyager is an open harness for creative work tuned so that AI models can work with creative tools more effectively.
Anthropic launched its Cyber Mission to apply frontier models, engineering support, and funding to critical-infrastructure and open-source security.
Sonnet 5.5 now runs around 20% cheaper on most agentic work.
Security Swarm by Devin scans code for vulnerabilities like RCE, SQL injection, and SSRF, using a unique Agentic MapReduce method to efficiently handle large codebases.
After training on 110 legal tasks, Harvey's wake-sleep agent improved held-out all-pass rates from 2.9% to 15.7% by distilling graded trajectories into reusable lessons.
Ethereum researcher Justin Drake issued a call for "bunker mode," urging crypto holders to migrate assets to fresh addresses whose public keys remain hidden behind a hash before a potential ECDSA break. AI-driven mathematical progress, including OpenAI's recent mathematical superintelligence release and a series of proofs disproving long-held conjectures, may enable fast private key recovery on GPU clusters within months. Elliptic curves carry rich algebraic structure exploitable through AI-discovered techniques, and the absence of cryptographic breakthroughs among OpenAI's 722 published mathematical results raises the possibility of government-suppressed cryptanalysis. The Ethereum roadmap already targets hash-based cryptography with end-to-end formal verification, and Drake is pushing to accelerate those timelines, while advising oracles and L2 security councils to rotate ECDSA pubkeys with every signed message.
Standard Chartered's Singapore unit plans to offer institutional custody for selected crypto assets, stablecoins, and tokenized real-world assets by the end of the year, subject to regulatory requirements. It extends the bank's existing custody footprint in the UAE, Luxembourg, and Hong Kong, and follows its move to bring Zodia Custody, the institutional custodian it incubated with Northern Trust, in-house.
Polygon Labs announced its Open Money Stack now supports TRON, letting fintechs and remittance providers connect bank funds to USDT on TRON through one integration covering regulated ramps in 48 US states, programmable TRON wallets, and cross-chain routing between TRON and EVM networks. Polygon says it has processed over $3 trillion in payments volume, and the stack was assembled through its pending Coinme acquisition and its Sequence acquisition.
Reap Virtual Assets lets card projects make BTC, tokenized stocks and gold, yield-bearing assets, points, or even credit lines spendable by setting a fixed price or live price feed per asset and combining several into one card balance. The assets stay with the project while Visa settlement happens in USDC or USDT on the backend.
Vitalik Buterin warns that AI-accelerated mathematical research may erode the security of lattice-based cryptography (ML-DSA, LWE, and FHE) through the same mechanism that GNFS degraded RSA, where decades of optimization shrunk effective key strength by an order of magnitude, and projects that AI could compress 50 years of math progress into two years. Ethereum's lean roadmap has already pivoted to hash-only signatures (WOTS and SPHINCS+), dropping lattices, ML-DSA, Falcon, and lattice-based commitments from ZK proofs. Public-key encryption cannot be constructed from hashes alone due to long-standing mathematical constraints, which means secure messaging, VPNs, and anonymizing protocols face an open problem that extends well beyond blockchains. Buterin's practical recommendations include keeping funds in addresses that have not yet broadcast transactions, moving multisig confirmations offchain to avoid exposing signer public keys, and routing encrypted notes through offchain channels rather than publishing them onchain.
Robinhood, eToro, Webull, and Coinbase all launched agentic trading accounts in Q3 2026, with Robinhood reporting 100K+ users and Coinbase securing the first SEC registration for an agentic investment adviser. x402 processed 232M transactions totaling $54M in volume, ERC-8004 agent registrations doubled quarter-over-quarter to 500K across 22 networks, and distributed RWA value reached $38.6B in Q3. Twenty-one major banks, including Goldman Sachs, Bank of America, and Citi, committed to forming a stablecoin company, Mastercard launched Agent Pay for Machines with 30+ partners, including Coinbase and Stripe, and Visa cited x402 and MPP as better suited to machine-native micropayments than cards. Singapore's MAS released the SAFR framework for AI agents in finance co-authored by Circle, Visa, HSBC, and JPMorgan, while the US CLARITY Act failed a Senate vote, leaving the SEC and CFTC to establish crypto regulatory rules.
Coinbase's head of cryptography Yehuda Lindell called Drake's warning that ECDSA could break within months "FUD," saying there is no evidence the hardness assumptions behind elliptic-curve cryptography have weakened. Dragonfly's Haseeb Qureshi called it "a very sober call," while Solana Foundation's Jacob Creech said Solana users do not need "bunker mode" because its Ed25519 setup could migrate to a quantum-resistant scheme through a hash-based proof.
Morpho's fee would have to come from suppliers (lowering yield on portable capital), curators (compressing margins and pushing them toward riskier vaults or other platforms like Aave V4), or both, unlike Uniswap, where liquidity benefits from the volume it attracts. Morpho still needs to prove pricing power, noting that Uniswap's recent strength came from an outlier month driven by Robinhood meme speculation, with volume now down about 71% from its September peak.
PropAMMs hold 5-10% of Ethereum DEX volume against 40% on Solana, with Fermi ($280M) and Metric ($178M) well behind Uniswap's $6B in the week of September 14-20.
Bitcoin fell to $82,800, breaking the technical level FxPro analysts identified as a trigger point below which sellers take control and prices could move toward $80,000.
BitMine held 6,016,414 ETH as of October 4, applying MicroStrategy's Bitcoin treasury accumulation model to Ethereum.
OpenAI's Intelligent UI lets ChatGPT combine text with interactive diagrams, charts, buttons, and task-specific tools instead of using the same format for every answer. Users can explore explanations, change inputs, or use a calculator directly within a conversation, with the layout adapting to the question. The feature is rolling out with GPT-6 across paid and free plans, and users can reduce visual elements when they prefer simpler responses.
Google's Nano Banana 2.1 reached the Gemini API for image generation and editing, at 1K, 2K, and 4K resolutions. Google reports better text rendering, prompt adherence, and character consistency across successive edits, plus fixes for tiling artifacts in wide and panoramic images at 2K and 4K. The model is available as gemini-nano-banana-2.1, and Google has deprecated gemini-3.1-flash-image without announcing a shutdown date.
ChatGPT has generated cartoons containing signatures associated with more than 15 New Yorker contributors, potentially making AI-generated work appear to have been created by real artists. Cartoonists including Brendan Loper and Emily Flake have encountered images bearing recognizable versions of their names despite having no involvement in creating them. OpenAI has introduced warnings for some New Yorker-style requests, but signed outputs have reportedly continued to appear, raising concerns that style imitation can cross into false attribution.
Slack's AI-drafted UI copy is almost indistinguishable from human work, and content designers are uniquely equipped to close the trust gap that "almost" opens. When Slackbot appeared to leak private information that was actually hallucinated, the team redesigned the UX to fetch correct data and explain why LLMs hallucinate. Three tenets guide the work: trust is the product, human judgment is irreplaceable, and language is infrastructure for AI agents.
An experimental general-purpose accessibility agent piloted at GitHub has reviewed 3,535 pull requests with a 68% resolution rate. It gives engineers real-time guidance in Copilot CLI and VS Code, and fixes straightforward issues before code reaches production. A parent agent directs two sub-agents, a passive reviewer and an active implementer, and escalates drag-and-drop, data grids, and rich text editors to human specialists.
A personal website redesign became an excuse to explore deliberately playful details, including fake game cartridges for past jobs, draggable photo prints, paper-like shaders, animated labels, unusual links, and dynamically generated social cards. Many of these touches serve little practical purpose, but that lack of optimization is precisely what makes the process enjoyable and gives the site its personality. Personal projects can create space for experimentation that professional work often removes, making unnecessary details, imperfection, and prolonged tinkering valuable in themselves.
Create physical products with Autonomyware: everyday objects, complex systems, sculpts, and toys, engineered autonomously from your idea.
A Safari extension for Mac that measures any website's colors, type, spacing, and layout, then hands it to your coding agent as one measured prompt.
Energy Icons is an open-source library of more than 1,200 icons for energy and climate products, covering areas such as renewables, electricity, transportation, buildings, and decarbonization in two optical sizes and two weights.
The M4 iPad Air's 12GB of unified memory brings it much closer to the iPad Pro for demanding creative work, comfortably handling large Procreate files, 3D models approaching 20 million polygons, and ProRes video editing. Its M4 chip delivers roughly 42% higher single-core and 44% higher multi-core Geekbench performance than the M2 Air, while Apple Pencil Pro support provides the same drawing experience as higher-end models. The iPad Pro still offers advantages such as an OLED display, nano-texture glass, and more memory, but many artists, illustrators, 3D artists, photographers, and video editors may no longer need to pay the premium.
Mother Design has created a restrained packaging system for thousands of M&S Home and Lingerie products by removing competing sub-brands and hierarchies and looking to classic information design rather than other retailers for inspiration. The system standardizes MS London typography, introduces a geometric numeral set called MS London Geo, and uses grids, consistent hierarchy, restrained colors, and color-coded sizing to make complex product information easier to navigate. Packaging materials, photography, and tone adapt between Home and Lingerie while maintaining a shared structure designed to strengthen the M&S master brand and scale into additional categories.
Spanish artist Javier Demsky, who started in graffiti in the early 1990s, bends the letters of his name into luminous structures blending typography, architecture, and science fiction.
Evaluating an information architecture starts with its scale, whether organization, platform, site, interface, or interaction, since only IAs at the same scale can be compared.
Sponsored
The GoDaddy Developer Platform connects hosting, domains, and professional email in one programmable workflow, accessible from your code, scripts, or AI agents. Deploy Node.js apps through APIs or the gddy CLI. Then use the APIs to register a domain, configure DNS, and set up professional email. One platform identity, consistent API conventions, and shared documentation connect each step. Access developer pricing on eligible domain registrations.
A reverse-engineering tour traces ChatGPT's interactive responses from a model-written DIL format through server-side compilation, sandboxed client execution, and native rendering. It also explains how streaming recompiles partial output, preserves keyed state, and sends server-resolved data such as images separately from model text.