Full-text search across article titles and summaries. Refine with sector, year, content type, and reading-time filters.
Sponsored
Your customers don't want to waste time. Parloa's AI agents eliminate holds [13] by managing millions of conversations, so your support team can focus on higher-value relationship building. Take for example, Decathlon, they eliminated 20% of their repetitive tasks with Parloa.
OpenAI plans to develop an automated AI researcher by March 2028, aiming to enhance research efficiency while maintaining human oversight to ensure alignment and safety. Researchers now use coding agents more frequently, with increased code generation and experiment execution, shifting focus to more complex tasks. The organization paused reinforcement learning training temporarily following a security breach but continues to adapt safety measures and transparency to uphold the development of safe AGI.
LLM-as-a-Verifier is a general-purpose framework that provides fine-grained feedback for any agent without requiring additional training. It achieves SOTA performance across coding, robotics, and medical agentic benchmarks. The feedback generated from the framework can be used for test-time scaling, progress tracking, and reinforcement learning.
Random Attention kept a uniformly sampled subset of generated KV-cache entries instead of relying on learned importance signals or attention statistics. Across several reasoning benchmarks and model families, it matched or exceeded more complex eviction methods while reducing eviction overhead.
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].
Anthropic is expected to complete its IPO listing days before the US midterm elections in November. It will begin marketing the offering in mid-October at the earliest. The IPO prospectus will likely be released in late September. The plans could change - such changes are not unusual, as companies frequently have to adjust schedules due to market conditions, regulatory reviews, and other preparations.
OpenAI claimed it had achieved AGI due to its 99.9% score on ARC-AGI-3. However, tests that ran the same model through the benchmark's own software scored 62.7%. The gap comes from the software around the model that OpenAI built. The different scaffolding around its agents helped OpenAI achieve the high score.
OpenAI President Greg Brockman discussed Astra, OpenAI's new model, focusing on its enhanced capabilities and alignment. He highlighted the importance of scaling infrastructure and addressed challenges in cybersecurity following the Hugging Face incident. Brockman shared insights on OpenAI's positioning within the tech value chain, emphasizing strategic focus on sectors like health and collaboration with partners like Nvidia and Microsoft.
The US data center capacity will expand from 25 to 70 gigawatts, requiring $5 trillion, mostly financed by debt. This expansion creates a 34% growth in the US corporate bond market and raises questions about financing, potentially involving municipal bonds. To service this debt, annual AI revenue must grow from $150 billion to at least $1.2 trillion by 2030, requiring a 55% annual growth rate.
Extropic unveils the Z1 chip to improve energy efficiency in transformer inference by leveraging probabilistic sub-threshold CMOS technology.
OpenAI knew about the message boards scattered across the internet that its agents created before the Hugging Face attack.
World Labs' Atlas unifies generation and 3D reconstruction through new-view prediction, using sparse images to infer scenes from unseen positions.
Mark Zuckerberg reportedly raised concerns about a national AI regulator during a call in August, saying that any appointees to the body should reflect the president's own light-touch approach.
AIRA₃ is a new generation of Meta's autonomous AI research engine that runs and coordinates many long-running agents asynchronously in their own isolated environments.
Google is advancing its Gemini desktop app with new features like Ask and Assign modes for enhanced functionality and potential remote control features.
Coinbase filed SEC-notice registrations for a derivatives exchange and broker, positioning the firm to offer single stock perpetual futures to US traders under formal regulatory oversight. The dual-regulator coordination with both the SEC and CFTC marks a deliberate effort to pull a crypto-native derivative structure into the US onshore market, expanding the range of instruments available through a regulated US venue.
~4,000 BTC was withdrawn from the Bitcoin sidechain's federation wallet by purported white-hat hackers via the SideSwap Peg-out Authorization Key, which Liquid says was not itself compromised. An on-chain message read, "We are whitehats, contact us on chain." Blockstream paused the network, exchanges suspended LBTC deposits/withdrawals, and other Liquid-issued assets were unaffected. No confirmation yet on whether funds will be returned or a restart timeline.
A new Solana primitive called payment channels addresses the per-call signature bottleneck in agentic micropayment workflows by letting agents authorize a spending ceiling once, meter usage off-chain via signed messages at zero transaction cost, and settle the full session in a single on-chain transaction across the x402 and MPP protocols. Funds remain non-custodial throughout, held in program-controlled escrow rather than operator balances, unlike prepaid credit systems where balances are tracked off-chain by the service provider. Benchmarks showed 100,000 unique wallets processing over 1 million payments per second, equivalent to more than 80 billion payments in 24 hours at a processing cost of $0.0078 per million payments. The spec is public, and the program is open source. Alibaba Cloud launched live inference endpoints with payment channels on day one.
Longbow is a new coin creation platform on Robinhood Chain that lets users launch tokens paired to any stock or crypto asset of their choice. The tool extends Robinhood Chain's utility by enabling user-generated linked instruments that bridge retail equity and on-chain token markets.
Orthogonal has scaled its x402-based API marketplace to 700+ endpoints across 50+ providers, with commercial workflows in contact enrichment, investor signal monitoring, and lead generation already processing real agent transactions. Building production-grade agent commerce on x402 rails required three layers the protocol does not itself provide: refund management, request-time dynamic pricing, and natural-language discoverability through the Bazaar catalog. An early investor-signals deployment grounded 87% of surfaced company updates within two weeks, up from near zero prior to the agent-driven approach. Orthogonal's near-term roadmap targets governance and spend controls, self-service API onboarding, and an enterprise control plane for managing agent capabilities across providers.
Aggregate blockchain throughput has grown over 100x in five years, with some production systems now clearing tens of thousands of TPS, removing raw capacity as the primary barrier to onchain finance. The conversation has shifted to market microstructure: millisecond inclusion uncertainty forces market makers to widen spreads against stale prices, and discretionary block-building gives MEV extraction the same structural character as privileged order-flow access in traditional markets. a16z researchers propose "Strong Chain Quality" to prevent single-actor control over block inclusion by reserving portions of each block for transactions surfaced through alternate network paths, while protocol designers are evaluating timelock and threshold encryption to hide transaction contents until settlement. Predictable execution, unambiguous priority rules, resilient access, and pre-execution confidentiality with post-execution transparency together define what a settlement layer must deliver before onchain finance works at institutional scale.
BitGo has agreed to acquire NYDIG IF Holdings, the institutional trading arm of Stone Ridge's bitcoin-focused subsidiary NYDIG, for $42.5 million in cash and stock plus a $15 million earnout. The deal is structured as $7 million in cash and approximately $35.5 million in BitGo equity, meaning the target receives roughly 83% of the base consideration in stock, with total consideration reaching $57.5 million if earnout targets are met. The acquisition adds institutional trading infrastructure to BitGo's existing custody and settlement stack, positioning the firm to serve a broader slice of the institutional client workflow under one roof.
Block is assembling the pieces to become a full payment network rivaling Visa/Mastercard. It already has millions of Square merchants and 60M+ daily Cash App users, and its new "Neighborhoods" feature can route payments directly between Block-owned accounts via Cash App Pay, bypassing card networks and charging just 1% while still profiting. Future phases could extend Cash App Pay to non-Square merchants (leveraging fraud data from Cash App's user base) and eventually make it an official debit network accepting all cards, positioning Block as a genuine fifth payment network.
Coinbase's Visa debit card is custodial, linked directly to a Coinbase account, while Bitget Wallet's card is self-custodial, charges 0% on USD spend but 1.7% on non-USD transactions, and offers up to 3% "Assetback" in a choice of assets.
13F filings show 30 institutions held a combined $74.9M in Hyperliquid ETF exposure (~1.15M HYPE) as of June 30.
Harmony announced it will fully shut down its Layer 1 network, citing security threats from "state actors and AI agents," and plans to migrate its ONE token to Ethereum via a snapshot-based airdrop.
US spot bitcoin ETFs saw $986.9M in net inflows last week, the third straight positive week, with BlackRock's IBIT leading at $691.5M.
Most of what a coding agent does is I/O rather than reasoning, such as reading five files to answer a question about one method. Spotify routed that work to two declarative Portal modes, bulk-reader and code-writer, both running Gemini 2.5 Flash on an ephemeral runtime, and reserved the frontier model for problems that need it. A quarter of engineering leaders already spend $200 to $500 per developer per month on tokens, and some are past $2,000.
This post explains why Meta put a managed proxy tier in front of ZippyDB, its massive key-value store. The real lesson is that proxies can turn client-fleet sprawl into a controllable infrastructure boundary: connection fan-in drops roughly 97–98%, batching and coalescing happen across clients, excess traffic from noisy tenants can be selectively rejected, and failover and load-balancing policy can change centrally without waiting for a million clients to update.
Vortex is an open-source columnar format designed to stream training data from S3 to GPUs with minimal copying. Cascading lightweight encodings, layout-aware pruning, and zero-copy memory paths target CPU and NVMe bottlenecks, giving ML platforms a way to keep expensive accelerators fed without converting everything upfront.
Semantic layers define business metrics, context layers provide supporting knowledge, and ontologies describe entities and their relationships. Together, they help AI agents interpret data, but people still need to keep that knowledge accurate and current.
Nao's founders argue that AI is shifting data teams toward maintaining business context and enabling self-service analytics. They recommend starting with one focused use case, measuring results, and improving data models before investing in a full semantic layer.
AI analytics can fail when agents generate free-form SQL against inconsistent warehouse metadata. Governed systems should route repeated questions through verified semantic-layer tools, parameterized calls, and explicit refusal paths, reserving open-ended SQL for cases humans can inspect. Maintaining definitions is the hard part regardless of model upgrades.
A 100,000-row Postgres load took 41 minutes because each row became its own statement and transaction; COPY cut the same work to 11 seconds. Real batching removes network round trips, parse overhead, and WAL flush barriers, while driver flags and transaction chunking prevent bulk APIs from silently behaving row by row.
Polars 2.0 makes the streaming engine the default for lazy queries, aiming for much lower memory use and faster execution while accepting that row order is not guaranteed unless requested. It also tightens type coercion and concatenation behavior, and removes deprecated APIs so data mismatches fail earlier instead of hiding in production pipelines.
Apache Iggy is a Rust streaming engine with Kafka-like logs and consumer groups, optimized around thread-per-core execution and modern Linux I/O. It looks promising for low-latency single-node or simpler deployments, but Kafka still has the deeper production story for replication, exactly-once semantics, compaction, and multi-region operations.
Sail 0.7 pushes Spark-style intermediate state toward object storage with blocking shuffle and checkpoint support. It persists shuffle outputs as compressed Arrow streams so downstream task regions retry without cascading upstream failures, while checkpointing keeps iterative jobs from rebuilding long query plans. The design targets stateless workers, smaller clusters, and better resilience under preemption.
Text-to-SQL assumes users ask questions that a single query can answer, when in practice they bring problems that need a sequence of queries with reasoning over the intermediate results. The paper proposes data investigations as the replacement paradigm and presents D2, a prototype that autonomously searches, reasons over, and collects data to reach evidence-backed conclusions, benchmarked on a Murder Mystery dataset.
Warner Music Group turned a hand-edited anti-counterfeit tracking spreadsheet into governed Databricks tables without taking the spreadsheet away from operators. The pipeline handles shifting tabs, renamed headers, late edits, row fingerprints, and full-overwrite loads so executive reporting can trust a messy business interface that people actually use.
PostgreSQL 19 Beta 3 adds temporal SQL, idempotent upsert-and-fetch, IGNORE NULLS window behavior, concurrent repacking, DDL extraction helpers, SQL/PGQ graph querying, and stricter NULL handling.
DuckFlight lets PostgreSQL tools and Arrow Flight SQL clients read and write the same live DuckDB database without a separate service or copying data.
Adobe has named Anil Chakravarthy its next president and CEO, succeeding Shantanu Narayen on December 1 after his 18 years leading the company. The handover lands as fears of AI disrupting software makers push Adobe shares down 25% in 2024, 21% in 2025, and 18% this year. David Wadhwani, a widely tipped contender who ran the creativity and productivity business, is leaving alongside the announcement.
Solaris is the first Interface World Model, generating apps and websites frame by frame as a user interacts, with no code layer in between. A study of 250 participants preferred it over a coded interface in 61% of instruction-following comparisons and 71% of those judging natural behavior. Legible text, trust in generated content, and coherence across long sessions remain unsolved.
Leaked circuit diagrams suggest Apple's upcoming A20 Pro chip may feature a 7-core GPU, an expanded cache for its efficiency cores, and a faster memory interface, potentially delivering improved graphics performance and power efficiency. While the source has a reasonable track record, the claims are based on indirect analysis of motherboard layouts rather than the chip itself, so they should be treated as informed speculation until Apple officially unveils the iPhone 18 Pro.
AI is becoming a useful part of UX work, but it works best as an assistant rather than a replacement for human judgment, research, and design expertise. The strongest use cases are targeted ones—speeding up research planning and synthesis, prototyping, writing, ideation, workshops, and service design—while keeping real users, critical evaluation, and clear accountability at the center. As UX roles evolve, the most durable skills are shifting toward strategy, critique, taste, systems thinking, and the ability to curate meaningful context for AI rather than simply producing artifacts faster.
As AI agents move beyond responding to prompts and begin actively navigating context, modifying files, running workflows, and making decisions, interface design is shifting from specifying every user action to orchestrating how responsibility moves between people and systems. The key design challenge becomes deciding what an agent can do autonomously, when it should ask for confirmation, what must remain reversible, and which judgment-heavy decisions should always return to a human. In this model, designers increasingly shape the product's policy layer—behavior, permissions, constraints, defaults, and attention—while human value moves toward deciding what deserves oversight and what can safely be delegated.
Generative plugins and shaders in Figma gain Community and organization publishing, animated and interactive effects, downloadable code, and MCP server support. The features began as a vibe-coded WebGPU prototype, then moved to a real pipeline system once a rendering engineer rebuilt the sandboxed architecture. With two months until Config, the team dropped PRDs and crits, working from demos and code prototypes in a single Slack channel.
Causal is an infinite AI canvas for moodboards, notes, files, and links. Plan visual projects in one space, with an agent that reads the whole board. Free on Mac.
A list of tips, improvements, and best practices that quickly make your interfaces better.