Full-text search across article titles and summaries. Refine with sector, year, content type, and reading-time filters.
Sponsored
AI investments keep pouring in, but engineers know it's not delivering everywhere. Particularly in observability, slipshod implementation is leading to wasted hours and slower incident response. Datadog spoke with 120+ engineering leaders [5] to discover: * Where AI already delivers the most value in production, and where teams overuse it * How data quality is the key to the trust needed to effectively
* The areas where AI is reducing speed and quality
Apple plans to launch several devices developed through a partnership with LG as part of an ecosystem of devices that work with its new smart home hub. These include a doorbell, thermostat, an upgraded HomePod mini, and a fresh TV set-top box. The products, set to launch on October 13, will carry the LG brand. Most of the devices will include support for Wi-Fi, Bluetooth, Thread, and Matter.
OpenAI released hundreds of new findings spanning a wide range of topics including algebra, number theory, theoretical computer science, mathematical logic, and topology. These were solved with an internal model that has not been released publicly. There is debate on whether OpenAI's systems were employing creative thinking or simply completing the final steps of a proof after borrowing ideas from human mathematicians' work. OpenAI has provided summaries of how its models came up with the solution for 10 of the problems.
Anduril recently announced that it was awarded a contract worth up to $1.8 billion over five years to scale Lattice across the Army's Next-Generation Command and Control effort. The company plans to build the second of its Arsenal mega-factories in Baltimore. The mega-factory will be shipbuilding-focused and will, when operations begin in 2030, span over 2 million square feet of manufacturing space. The facility will be backed by $6.6 billion across both private investment and up to $2.9 billion through a submarine production contract with the US Navy.
Building nuclear power plants takes a long time, so some companies are looking to upgrade older reactors to squeeze even more power out of them instead. These types of upgrades can cost billions and require approval by regulators, but they can often be completed in five years or less at a fraction of the cost of building a new plant from scratch. Many tech giants are eager to pay for these upgrades to power AI. Increasing production from existing nuclear plants will add an estimated 6,000 to 8,000 megawatts of electric capacity to US grids.
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 [12].
The speed of change in the tech industry has never been this large-scale or fast. Practically nobody writes code by hand anymore. Assumptions about code output have broken, code reviews have become theatrical, and quality and reliability are down. Teams are still important, as is planning. Engineers will likely stop reading code altogether as companies start building a new type of AI infrastructure.
OpenAI's Decisions API turns text and images into decisions that applications can use. It evaluates text, images, or both and returns typed answers 10 times faster than the Responses API. The Decisions API returns a probability that a condition is true, a choice from a fixed set, or a score against a rubric. These answers can be used to classify content, route requests, and prioritize work in an application.
We used to learn things by doing something wrong first and then taking a lesson out of it. LLMs make everything so fast that we don't want to spend time doing the wrong things, even though it's more beneficial in the long run. The struggle is how the brain actually learns new things. Use AI to help you think through problems rather than just giving you the answer.
Decision models return a choice, a score, or a yes-or-no with a confidence number stapled on. Within a day of launch, TypeSafe's Jev decision model had been wired into 13% of Vercel's paid AI Gateway customer workflows, and within three days, Cloudflare, LangChain, and Langfuse had shipped integrations for it. It took competitors about two weeks to show up. The field is evolving quickly - this post looks at the criteria and nuances of using decision models in production to help readers decide which model to use.
Sponsored
Clarify is the CRM that works while you sell. Its agents prep meetings, update records, find prospects, draft outreach, and more. 1
Jiti is a small kernel for growing a running Lisp application through conversation with an LLM - users ask for a capability, the model writes Lisp, and the application permanently acquires that capability until the user asks for that capability to be removed.
The Stack Overflow Developer Survey 2026 asked developers and technologists about everything from how they're using AI to how happy they are at work to where they're learning new skills.
JPEG XL offers 30% to 50% better compression than JPEG, lossless compression, built-in HDR support, lossless JPEG transcoding, and more.
The DNS root is scheduled to change its key-signing key for the second time ever on October 11.
Codemode is a way for LLMs to express and orchestrate complex operations on the harness side, but not the execution environment side.
OSC 7501 is a specification for a terminal escape sequence that lets any program tell the terminal whether it's idling, working, waiting on the user, finished, or failed, and why.
Sponsored
Copilot. Cursor. Agentforce. Your teams switched on agents months ago. Ask who reviewed them, then what they can reach. That gap is now its own market: GARTNER'S EMERGING MARKET QUADRANT FOR AI APPLICATION SECURITY named Zenity one of two startup Market Shapers, after naming
🧭 ENTERPRISE AGENTIC AI SECURITY BUYER'S GUIDE: what to evaluate before your next deployment, from must-have controls to cost. DOWNLOAD
Mistral Large 4 is a one-trillion-parameter multimodal model positioned as a European alternative to leading closed and open AI systems.
Nano Banana 2.1 is a member of the Gemini 3 series of models. It can take text and image inputs and output images and text. Based on Gemini 3.6 Flash, the model has a token context window of up to 1 million. Nano Banana 2.1 is available in the Gemini app and API, Google AI Studio, Google Search AI Mode, Google Ads, Google Flow, and Google Stitch.
OpenAI has released a broad range of mathematical results produced by an internal frontier model. It has published the results in a GitHub repository along with protocols for paper revisions and citations. The repository contains formalizations of many of the proofs in Lean and will be updated with more formalizations as they are obtained. It also includes 10 summaries of the model's reasoning, estimations of compute spent in terms of Pro usage on ChatGPT, and statistics about the number of attempted problems.
Open-weight models potentially pose an untenable risk to society's functionality. Chinese companies continue to release open-weight models with strong cyber capabilities. GLM-5.3 crosses a threshold of capability, but there is little public evidence that much has changed. Some say that open-weight models will result in a crippling of cyber infrastructure - after many years of debating open-weight risks, we will finally start to get some real answers.
Google Project Zero examined how large software vendors can patch urgent vulnerabilities faster than their normal release cycles allow. The guide breaks down the full remediation pipeline (from triage and development through testing, delivery, and activation) and highlights systems designed for exceptional security incidents.
Sponsored
Models and agents ship faster than governance can track. Score your stack on 12 checks, from lineage to risk tiering, and find trust gaps keeping pilots out of production. Five minutes; answers stay in your
The Personal Agent Protocol is an open standard that Meta and Sierra are developing along with other industry partners. It is designed to handle authentication, empower consumers, and give companies visibility into what personal agents do through their websites, APIs, or company agents. Consumers decide what access to give their personal agents, and companies set parameters for what those agents can do. The protocol is open for anyone to implement.
OpenAI's Decisions API is now available to all developers in public beta. It makes decisions up to 10 times faster than GPT-6 Luna through the Responses API. The Decisions API accepts text and image inputs and supports three kinds of outputs: predicates, choices, and scores. It costs $0.10 per 1 million input tokens - there are no cache-read, cache-write, or output-token charges.
EmbeddingGemma 2 is a 740M-parameter model that maps text, code, images, audio, and video into a shared embedding space. Built for on-device inference and released under Apache 2.0, it enables multimodal search and retrieval without sending data to the cloud.
Pantheon rapidly assembled a robust UMI data collection operation, creating a diverse dataset to improve dexterous manipulation in robotics. They built custom infrastructure and hardware, drastically reducing data collection costs from $60/hr to $10/hr while ensuring data diversity. By incorporating innovative freeform and scripted methodologies, Pantheon accumulated over a million unique tasks, overcoming traditional data limitations for enhanced AI model training.
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 [18].
DeepSeek is nearing a $12 billion raise, surpassing its initial 50 billion yuan target, with Tencent and CATL as major contributors. The funding boost follows the release of DeepSeek's latest AI model, with expectations it might reach 100 billion yuan. The company is also preparing for a potential Shanghai STAR Market IPO, employing CITIC Securities for assistance.
China has made recruiting the world's best scientists a national priority. It introduced a visa for scientists and is offering lucrative research grants. Foreigners are not coming despite those efforts. The number of Chinese scientists working in the US has risen rather than declined.
Hark has released an AI agent that buys groceries, books cars, and pays bills.
Claude now integrates with Google Docs, Sheets, and Slides for all paid plans, allowing direct editing within files.
DeepSeek V4.1 Flash, a Chinese AI model, now leads in agentic coding benchmarks, surpassing US competitors.
NVIDIA's AI Cluster Runtime (AICR) v1.0 offers version-locked, validated recipes for stable GPU-accelerated Kubernetes cluster configurations.
Anthropic expanded its Cyber Verification Program into three access tiers for qualifying security professionals, providing its most capable models with reduced cyber blocking.
FinCEN is withdrawing its 2023 proposal that would have designated international crypto mixing a "primary money laundering concern" under Section 311, along with a 2020 proposal requiring banks and money services businesses to verify identities and keep records on self-hosted wallet transactions above $3,000. The agency said commenters' concerns about the mixing rule's broad definition chilling legitimate privacy use informed the decision, and because neither proposal was finalized, existing obligations do not change. FinCEN said it will keep monitoring mixers for illicit use and cited a July 2025 President's Working Group report acknowledging lawful users may use mixers for financial privacy.
Privy founder Henri Stern has taken on a dual role leading crypto and stablecoins at Stripe, tasked with building a stablecoin developer platform designed to be open, modular, and global by default. Stripe's existing client base shows the demand: Meta uses the rails to pay creators in real time, Ramp runs fiat and stablecoin flows in parallel for global expansion, and Deel routes contractor payouts worldwide, each requiring end-to-end orchestration rather than isolated point solutions. The platform integrates Bridge, Privy, Tempo, and Open USD together, treating fiat and crypto as interoperable layers rather than separate stacks businesses must rebuild from scratch. Stripe holds a conditional OCC national bank trust charter approval in the US along with MiCA and EMI licenses in Europe, placing it in a GENIUS Act-ready compliance posture.
Polymarket is replacing its Gnosis Conditional Tokens Framework (CTF) from 2019, which accumulated contract layers with each new feature, with Protocol V2: a ground-up rebuild that consolidates the stack into one ERC1155 positions contract, one pUSD collateral token, one exchange, and one router across all market types. A new OracleAggregator with pluggable resolution modules connects to UMA, Chainlink, and future oracle sources, while first-class bridging support for positions, collateral, and resolutions lays the infrastructure for multi-chain expansion. The protocol launches with four market types (Binary, Atomic Neg-risk, Incremental Neg-risk, and Combinatorial) and is upgradeable behind a governance process, enabling the team to ship features without disrupting existing integrations.
Zcash activated its NU7 upgrade on the public testnet at block 4,465,026, cutting target block time from 75 seconds to 25 seconds and reducing three-confirmation finality from 225 seconds to 75 seconds. The upgrade splits fees 40/60 between miners and a reserve for future mining rewards, and reduces per-block rewards to one-third their current amount to maintain ZEC's issuance schedule despite the tripled block rate. NU7 also retires Sprout, Zcash's original shielded payment protocol, requiring holders to migrate funds before the planned November 5 mainnet activation. Both major node implementations, Zakura and Zebra, shipped compatible releases in the first two days of October, with developers targeting an October 20 decision on mainnet confirmation.
Chainalysis attributed the September 24 Bitget breach to North Korean actors, a $387M theft that pushed the DPRK's 2026 crypto total past $1 billion. Within three hours of the hack, attackers dispersed funds across Ethereum (49.7%), XRP (40.8%), Zcash (7.6%), and Tron (1.8%) in 23 transfers. Custom AI tools compressed what would have required 20+ hours of manual bridge reconciliation into under 10 minutes, with human investigators directing the analysis throughout. Attackers routed stolen XRP through cross-chain liquidity protocols to convert it into Bitcoin over roughly 36 hours before funds reached attacker-controlled addresses, avoiding direct exchange exposure.
Near Intents recovered its full $3.8 million in stolen funds within 48 hours after General Manager Alex Shevchenko identified the attacker on-chain and issued a return deadline with an implicit deanonymization threat. The exploit targeted a flaw in how the platform's Omni deposit/withdrawal layer interacted with its main smart contract, forcing Near Intents to halt operations across its 35-blockchain cross-chain swap service. The attacker returned all funds and posted an on-chain message acknowledging fault and urging others to use bug bounties rather than exploits. The incident came days after Near Intents blocked a $50 million swap attempt linked to the $387.5 million Bitget breach, marking the platform's second high-profile security event within a week.
Ondo Private Markets will offer tokenized notes tied to the per-share value of selected private companies at qualifying liquidity events, starting with a pre-IPO AI firm whose notes are expected to begin 24/7 secondary-market trading this week. The notes are not shares and carry no ownership or shareholder rights, are obligations of the issuer, and are available only to eligible non-US investors, with robotics, cybersecurity, biotech, and infrastructure companies to follow. Ondo said its tokenized stocks and Treasurys platforms hold $3.7 billion in TVL with more than 1 million cumulative holders, while tokenized equities overall exceed $42 billion in market cap, about 35% of the $121 billion tokenized-asset market.