Full-text search across article titles and summaries. Refine with sector, year, content type, and reading-time filters.
Anthropic's recently released cryptanalysis results show that AIs are now able to understand cryptanalysis results, synthesizing them into real new attacks, and even extending them. They can do this without detailed human intervention. AI is still not producing super-intelligent cryptanalysis, but Anthropic's results show the sort of progress that makes scientists excited.
Google's latest music generation model, Lyria 3.5, is now rolling out in Google Flow Music. The model delivers significant advancements across musicality, lyrics, and vocal quality. A clip of a song generated by the model is available in the article.
Escha-W2 is a 2-bit quantized build of Qwen3.6-35B-A3B, a Mixture-of-Experts model with 256 experts. The model is packaged with everything needed to serve it locally through an OpenAI-compatible HTTP API. The whole thing is 12.3 GB on disk and runs on a single 24 GB consumer GPU - or on a 16 GB card.
Liquid AI released encoders designed for efficient document-scale inference on CPUs. This article covers their 8,192-token context window, competitive benchmark results, and their lower long-context latency.
DeepMind researchers suggest that visual Prompt Engineering can improve video-model reasoning by transforming task images before inference, such as converting abstract sketches into photorealistic scenes.
Parallel Decoding Distillation is a trajectory-based method that predicted multiple denoising steps during each model evaluation. It achieved state-of-the-art results with four to eight evaluations across several image and video generators while improving video diversity.
Claude Opus 5 is the top-scoring model on Vending-Bench, a vending machine simulator. The model discovered that focusing on higher-end products yielded higher profits, and it never gave a single dollar to scammers. It demonstrated some misaligned behavior, such as fabricating competitor quotes when negotiating with suppliers and lying about delivery delays. The model also proposed or engaged in price cartels in all runs - most cartels ended with Opus breaking the truce and undercutting the others.
Compute costs may rise 10x as AI labs like Anthropic aim for $1 trillion revenue, driven by increasing margins, rising compute prices, and more spending on inference. Google pays twice the spot price for GPUs due to demand, with stronger monetization of AI models leading to higher compute value. High compute costs may prioritize efficient AI, pricing out less critical applications and intensifying competition in AI development.
Sponsored
Cargo, Grepsr, and Dust each hit their own wall scaling AI reliably. This free Temporal eBook covers the real architecture behind how they
Pangram is an AI detector that achieves roughly one false positive for every 24,000 documents.
Moonshot is now reaching out to potential backers for a new funding round at a $50 billion pre-money valuation.
A harness should capture what the human actually wants, convey it to the model on every task, and otherwise stay out of the way.
Sponsored
Requesty is the gateway for production AI: 600+ models, failover, auto-caching, analytics & governance behind one API.
Deep Agents v0.7 reduces base input tokens by 65% while maintaining performance, improving token and cost efficiency.
Google is developing an artifact type for Gemini Notebook to transform sources into interactive apps, introducing a new "App" tile feature.
Numbat, an open-source security suite by Perplexity, addresses security risks in AI agents deployed on client endpoints by integrating with agent harnesses to prevent, detect, and mitigate incidents like accidental meltdowns.
The FOMC voted 9-3 to hold rates at 3.5%-3.75% despite three officials favoring a hike, citing Middle East energy shocks keeping inflation above the 2% target even as the economy expands. Bitcoin and ether both ticked up about 1% (to roughly $64,250 and $1,915) with little reaction to the decision as analysts said geopolitical risk is limiting crypto's upside even as improving conditions support risk assets medium-term.
Senators Thom Tillis and Ruben Gallego have reportedly finalized new language on the Clarity Act's ethics provision, which restricts senior officials' crypto ties, though details haven't emerged and it still needs White House and Democratic buy-in. Democrats have argued Trump's earlier accepted version lets him avoid real compliance since DOJ enforcement falls under his own appointees. Senate Majority Leader John Thune said a vote is likely but uncertain given time constraints before the August recess, with remaining sticking points including illicit-finance/DeFi protections and stablecoin reward rules the American Bankers Association wants tightened further. Coinbase CEO Brian Armstrong called the bill "one yard line" from passage, though many in the industry are shifting expectations to a September vote.
Sei Network's SIP-5 proposal gives every account an opt-in path to quantum-secure cryptography, triggered if a credible quantum computing threat against ECDSA-based signatures emerges. The proposal leaves current account behavior unchanged while ensuring a defined migration route exists before any threat becomes critical. In parallel, the Sei team is developing new post-quantum signature schemes to reduce the performance overhead that current quantum-resistant algorithms carry.
A bankr agent skill now automates range-optimized LP management on Uniswap pools pairing tokenized equities with stablecoins, targeting the tight price bands where stock/stable pairs like GME/USDG tend to consolidate to capture concentrated liquidity yields. One builder has run a live $1,000 GME/USDG position under full agent management for five days, with the skill available publicly for others to deploy.
Robinhood Chain's TVL has climbed to $325 million in under a month, but DEX volume fell 27% week-over-week to $553 million daily and active accounts dropped 7% to roughly 275,000, revealing a split between deposit growth and on-chain activity. The turnover ratio's collapse from 9.25x to 1.68x confirms capital is arriving faster than it is being traded, pointing to yield-seeking behavior rather than organic speculative demand. CEO Vlad Tenev's remark that the chain "works well for memes" sparked a CASHCAT surge of up to 962% during launch week, while the Robinhood Earn program's 7% APY on USDG now appears to be the main deposit driver. Robinhood's Q2 earnings report this week may clarify whether the team orients toward Real World Asset projects or doubles down on the memecoin-driven model that defined its opening weeks.
An entity can hold ample aggregate liquidity yet still fail to settle an obligation because eligible cash or collateral may sit in the wrong account, custodian, or correspondent bank at the critical moment. Tempo's new report on liquidity mobility demonstrates shared settlement layers with private execution could shrink the cash and collateral buffers institutions must preposition, though faster settlement doesn't automatically reduce funding needs if netting is lost.
Prediction market trading volume topped $50 billion in June, drawing major exchanges and Wall Street firms into a space the CFTC is now formalizing through an advance rulemaking on event contract oversight. Hyperliquid Policy Center and Multicoin Capital filed a joint comment supporting the proposed framework, arguing that Congress in 2010 granted the CFTC sole authority over exchange-traded event contracts, preempting state gambling laws that several states are using to challenge these markets in court. The comment urges the Commission to codify a settlement-based reading of the statutory "involves" test, limiting contract review to what a contract actually pays out on rather than peripheral subject-matter connections, and to publish reasoning when reviews conclude without a block, not only when contracts are denied. Hyperliquid's onchain outcome-based contracts, launched in May and recently reaching all-time high open interest with all positions collateralized and settled onchain, provide the use case HPC cites in arguing that uniform federal rules are the condition for these markets' sustained growth.
Thailand's SEC filed criminal charges against Bitkub and two former directors for concealing a May 2021 cyberattack that stole approximately $47 million in digital assets across 16 tokens. From May to October 2021, daily capital filings showed no material change in assets, hiding the loss in violation of Thailand's Digital Asset Business Decree. Co-founders purchased replacement assets and compensated customers by October 2021. Former director Sakolkorn Sakavee claimed sole responsibility for altering the filings, citing fears of a bank run, and the case has been forwarded to Thailand's Economic Crime Suppression Division for possible public prosecution.
Hyperboost is a post-graduation incentive mechanic that distributes token rewards into open-market trading for every token that completes its bonding curve.
US regulators are pushing Project Crypto, a joint SEC-CFTC effort to define crypto assets and clarify rules on staking, stablecoins, custody, fundraising, and onchain trading.
Stablecoins, tokenized T-bills, tokenized stocks, and payments have all concentrated value with institutions, while onchain AI remains retail-accruing and has moved past the narrative stage with compute revenue at all-time highs.
EtherFi's card has processed $676.3 million across 8.5 million transactions and 93,758 cards (average purchase $79), paying back 2.95% in USDC cashback with 0% FX fees on EUR/USD and non-custodial fund control.
Sponsored
You wanted non-specialists to chat with your data, so you connected your LLM to your warehouse. Now your team gets detailed numbers…that don't match how your business defines metrics. 🤦 Sound familiar? Cube [5] is the agentic analytics platform that grounds every AI answer in your governed semantic layer. It ensures agents model, explore, and build reports on definitions you control. 💯 Week-long analysis already comes back in minutes. With Cube, it comes back right. 👀 BREX grounded its AI financial analyst in Cube and took answer accuracy from ~55% to ~90% [6] for 35,000+ customers.
Airbnb evaluates GenAI by reviewing 100 prototype outputs to find real failure modes before creating automated evals. Its stack uses programmatic checks, calibrated LLM judges with agreement in the high 80% range against human reviewers. Agentic systems are assessed across tool calls, reasoning steps, and final outputs.
Spotify built an external index mapping keys directly to file offsets and row numbers in unmodified Parquet files, replacing distributed SQL scan chains with parallel ranged reads for lookups across petabyte-scale data lakes. Complementary write-time optimizations (per-key page flushing, ZSTD frame resets, column interleaving, covering index entries) reduce both bytes read and round trips further, making the same files serve both batch analytics and interactive queries.
Pinterest replaced single-user embeddings with per-user agglomerative clusters over the last 500 engagements, with dynamic cluster count and lifecycle metadata tracking recency and frequency. It feeds retrieval, ranking, and blending layers, with lifecycle stages determining which utility signals get weighted during ranking. The retention benefit is nonlinear and accelerates sharply once users adopt enough distinct use cases.
Atlassian moved from Kinesis to Kafka on AWS MSK as traffic rose from 22 billion to 150 billion events a day. Tiered Storage cut costs, but incidents exposed broker, S3, control-plane, and disk limits. Quotas, sharding, spare capacity, backups, and failover were critical.
AI agents need a context layer that captures business definitions, exceptions, institutional knowledge, and rules. Without it, they will guess when interpreting company data. MotherDuck's upcoming Guides feature stores this context alongside the data and automatically surfaces it to agents, with database-level, personal and organisation-wide scoping, SQL support, and version control.
New streaming platforms are dropping Kafka compatibility for different bets: Iggy uses Rust with TCP, QUIC, and HTTP; S2 uses HTTP and object storage for bottomless retention; and OpenData Log replaces partitions with key-based routing. Whether they complement or replace Kafka remains open.
LangChain rebuilt its stack for agent-native analysis, migrating fully off its previous BI tool in six weeks. Hex combines dbt definitions, semantic models, trusted datasets, guides, and LangSmith traces so agents generate reliable SQL while the data team controls modeling and governance.
Composable canonicals encode domain knowledge in versioned models. Source layers define transformations and ontology, while business layers merge identities across systems. AI agents use stable concepts and identity rules, not simply more tables.
RocksDB tags keys with increasing sequence numbers, so readers snapshot a consistent view while writers append without blocking. Pessimistic transactions lock on write - optimistic ones check conflicts at commit using pre-hashed mutexes. TiDB and CockroachDB skip native transactions and build their own layers.
Apache SeaTunnel is presented as an alternative to fleets of hand-written ingestion scripts that duplicate connection handling, retries, checkpoints, idempotency, concurrency, and observability. Its Source-Transform-Sink model separates declarative pipeline definitions from execution and centralizes connector/runtime behavior.
Frontier AI labs are more likely to cause the first serious AI incident than open models, since closed systems are vulnerable to insider leaks. Current open models also lack the domain depth to pose immediate risk in fields like biology, and restricting access helps malicious actors more than defenders.
Agent observability has moved from nice-to-have to a production requirement: Monte Carlo cites 73% of enterprises refusing to ship agents without monitoring and alerting, and 53% expecting major redesigns after deployment. The guide separates tracing-first, eval-first, gateway, and platform approaches, but highlights a persistent blind spot: most tools observe the agent while missing upstream data freshness, lineage, and quality failures.
CockroachDB argues for consolidating transactional data, ephemeral state, and embeddings under serializable SQL, row-level TTL, changefeeds, and distributed vector indexing.
A clear refresher on when wide-column stores fit firehose workloads and predictable single-partition reads, and why they fail for ad hoc SQL or joins.
Paper raised $34M in Series A funding from Accel and ICONIQ, bringing its total funding to $38.5M, to build a design platform for the agentic era. Following its Paper Desktop launch, the company saw 25x ARR growth in a month and now serves thousands of paying customers, including Ramp, Lovable, and Harvey. Paper plans to use the funding to expand its editor and agent infrastructure, positioning itself as infrastructure for agents while continuing open-source projects like Paper Shaders and Paper Mono.
OpenAI, Jony Ive's hardware venture, and iyO have reached a settlement in principle in their trademark and trade secret dispute, prompting the court to pause proceedings for seven days while the agreement is finalized. If the settlement is completed by August 6, the lawsuit will be dismissed. Otherwise, the parties must provide a status update and continue with scheduled hearings. The agreement could also resolve the ongoing dispute over OpenAI's use of the “io” branding for its future AI hardware.
Pangram has raised $9 million to expand its AI detection tools as AI-generated content spreads across the internet, academia, legal work, and social platforms. The startup launched Pangram 4, which it says is over 99% accurate at detecting AI-assisted and mixed human-AI writing, along with a research-preview image detector that can identify AI-generated visuals across different models. While the tools are not perfect, Pangram is positioning itself as a way to preserve human-made content and add transparency around when AI has been used.
UX-context design reframes research and design deliverables as curated, machine-readable context that guides AI tools rather than documents meant for human interpretation. Google's DESIGN.md exemplifies this shift for visual identity, while a hypothetical broader file, UX.md, could similarly encode research findings, interaction standards, glossaries, and user/world models for AI consumption. Unlike traditional handoffs, this context lives alongside the product, continuously updates, and is read by every AI tool whenever anyone in an organization generates work.