Development & IT Daily Brief
20 events20 source observations
Lead stories
How we raised Databricks cluster rate limits by 7.5× | Databricks Blog
Databricks raised default cluster-creation request-rate limits by 7.5× and removed over a hundred customer-specific overrides in the reported scope.
Read the analysisThe day in context
Databricks raised default cluster-creation request-rate limits 7.5×, using adaptive concurrency while retaining a fixed guardrail. Google Cloud’s OKF v0.3 adds interoperability and trust rules without breaking existing bundles. AWS introduced HyperPod workload recovery and checks, and described Salesforce routing inference across its existing regional endpoints. Google Cloud highlighted European startups at Slush 2026. JetBrains says its updated LinkedIn plugin keeps usage data local; AWS’s Aurora chatbot design implements explainability in depth, not all eight responsible AI dimensions.
JetBrains recommends Native WSL mode for WSL projects from RubyMine 2026.3: the IDE runs on Windows while an agent handles Linux files and processes. PostgreSQL users have pgjumbo, a free cross-platform client, and pgSherpa, a free macOS monitoring companion. WAL-G 3.0.10 lets standbys read archived WAL from object storage. Apple’s HomeView 27.2.1 fixes two Siri issues and a home-hub issue. InfoQ reports that Vercel’s open-source UI SDK composes interfaces from reusable behaviors and remains in beta.
TechRadar reports that research unlocked more memory and compute on Nvidia’s CMP 170HX, but the A100 comparison rests on one benchmark. The US expanded AI chip rental controls; China said it would take countermeasures. Apple released Safari Technology Preview 257. Researchers found two versions of Microsoft’s official Python agent framework hijacked; affected users should assess account security. Fourth-generation Siri Remote firmware was updated without release notes. TechSpot and WIRED also report restrictions on overseas remote computing and advanced-chip shipments to five countries.
Lead stories
How we raised Databricks cluster rate limits by 7.5× | Databricks Blog
Published:
Databricks says it raised default cluster-creation request-rate limits by 7.5× and removed more than a hundred customer-specific overrides in the reported scope. Creating a cluster on its classic compute platform is a long-running workflow: after the API accepts a request, background processes still coordinate services and provision cloud resources. A fixed request rate cannot respond to unfinished work or changing system health. It can reject requests when capacity is available, yet continue accepting them when dependencies slow down. Databricks instead uses adaptive concurrency limits to adjust how much work starts as capacity and health change. Bounded queues and fair scheduling help absorb bursts and share capacity, while a fixed API rate ceiling remains a final guardrail. For customers planning cluster launches, the account explains the basis for the higher default limit; it does not promise constant capacity or launch speed, because system conditions still affect concurrent progress.
How we raised Databricks cluster rate limits by 7.5× | Databricks Blog
Published:
Databricks says it raised default cluster-creation request-rate limits by 7.5× and removed more than a hundred customer-specific overrides in the reported scope. Creating a cluster on its classic compute platform is a long-running workflow: after the API accepts a request, background processes still coordinate services and provision cloud resources. A fixed request rate cannot respond to unfinished work or changing system health. It can reject requests when capacity is available, yet continue accepting them when dependencies slow down. Databricks instead uses adaptive concurrency limits to adjust how much work starts as capacity and health change. Bounded queues and fair scheduling help absorb bursts and share capacity, while a fixed API rate ceiling remains a final guardrail. For customers planning cluster launches, the account explains the basis for the higher default limit; it does not promise constant capacity or launch speed, because system conditions still affect concurrent progress.
LinkedIn Connected Apps 1.2: Clearer Level System, Visible Progress, and Direct Answers About Privacy - The JetBrains Blog
Published:
JetBrains has updated its LinkedIn Connected Apps plugin, which tracks how developers use their IDE, assigns levels, and offers a level description they can choose to share. Version 1.2 reduces the system from five levels to four and introduces separate Coding and AI branches. Everyone starts at Level 1 and completes the first two Coding levels before the AI branch begins at Level 3; the two branch levels can differ. A widget shows the current level, requirements for the next one, and progress. On privacy, JetBrains says usage events that count toward a level stay on the computer and are not shared with LinkedIn. After an account is connected and a new level is reached, the IDE sends only the corresponding statement. Sign-in uses LinkedIn’s authorization page; the plugin does not see or store the password, and tokens are kept in the IDE’s secure credential store.
LinkedIn Connected Apps 1.2: Clearer Level System, Visible Progress, and Direct Answers About Privacy - The JetBrains Blog
Published:
JetBrains has updated its LinkedIn Connected Apps plugin, which tracks how developers use their IDE, assigns levels, and offers a level description they can choose to share. Version 1.2 reduces the system from five levels to four and introduces separate Coding and AI branches. Everyone starts at Level 1 and completes the first two Coding levels before the AI branch begins at Level 3; the two branch levels can differ. A widget shows the current level, requirements for the next one, and progress. On privacy, JetBrains says usage events that count toward a level stay on the computer and are not shared with LinkedIn. After an account is connected and a new level is reached, the IDE sends only the corresponding statement. Sign-in uses LinkedIn’s authorization page; the plugin does not see or store the password, and tokens are kept in the IDE’s secure credential store.
Design responsible AI controls for Amazon Aurora PostgreSQL | Artificial Intelligence
Published:
AWS describes responsible AI controls for a customer-support chatbot built on Amazon Aurora PostgreSQL-Compatible Edition. The chatbot retrieves historical tickets and knowledge-base articles to produce context-aware answers, but sensitive customer data raises privacy, accuracy, safety, and governance concerns. In the reference architecture, an ingestion pipeline redacts personally identifiable information and generates vector embeddings. Queries then use semantic search, Amazon Bedrock Guardrails, model inference, and interaction logging. AWS also describes row-level security and citations as controls, and says Aurora ML can invoke models from SQL to generate embeddings without a separate data-movement pipeline. The post maps eight responsible AI dimensions to relevant capabilities but implements only explainability in depth; it is not a complete implementation of all eight. For teams designing retrieval-augmented customer support, the example provides a way to think through citations, access controls, and logs, rather than a finished, comprehensive compliance solution.
Design responsible AI controls for Amazon Aurora PostgreSQL | Artificial Intelligence
Published:
AWS describes responsible AI controls for a customer-support chatbot built on Amazon Aurora PostgreSQL-Compatible Edition. The chatbot retrieves historical tickets and knowledge-base articles to produce context-aware answers, but sensitive customer data raises privacy, accuracy, safety, and governance concerns. In the reference architecture, an ingestion pipeline redacts personally identifiable information and generates vector embeddings. Queries then use semantic search, Amazon Bedrock Guardrails, model inference, and interaction logging. AWS also describes row-level security and citations as controls, and says Aurora ML can invoke models from SQL to generate embeddings without a separate data-movement pipeline. The post maps eight responsible AI dimensions to relevant capabilities but implements only explainability in depth; it is not a complete implementation of all eight. For teams designing retrieval-augmented customer support, the example provides a way to think through citations, access controls, and logs, rather than a finished, comprehensive compliance solution.
Complete briefing
Cloud & infrastructure
4Introducing OKF v0.3 | Google Cloud Blog
Published:
In June we introduced the Open Knowledge Format (OKF), an open specification that provides plain markdown and YAML to give agents the context they need. In July, v0.2 added trust signals: provenance, trust, freshness, lifecycle, and attestation.
Introducing OKF v0.3 | Google Cloud Blog
Published:
In June we introduced the Open Knowledge Format (OKF), an open specification that provides plain markdown and YAML to give agents the context they need. In July, v0.2 added trust signals: provenance, trust, freshness, lifecycle, and attestation.
Introducing workload recovery and expanded infrastructure checks on Amazon SageMaker HyperPod | Artificial Intelligence
Published:
Today, we’re introducing workload recovery and infrastructure checks on Amazon SageMaker HyperPod that catch these failures sooner and raise the goodput of your training workloads.
Introducing workload recovery and expanded infrastructure checks on Amazon SageMaker HyperPod | Artificial Intelligence
Published:
Today, we’re introducing workload recovery and infrastructure checks on Amazon SageMaker HyperPod that catch these failures sooner and raise the goodput of your training workloads.
How Salesforce is advancing global-scale model serving with cross-Region inference routing in Amazon SageMaker AI | Artificial Intelligence
Published:
Thanks to Amit Gupta, Senior Director- AI/ML and Mark Wang, Software Engineering PMTS from Salesforce for your contributions in the vision and partnership of cross-Region inference routing on SageMaker AI.
How Salesforce is advancing global-scale model serving with cross-Region inference routing in Amazon SageMaker AI | Artificial Intelligence
Published:
Thanks to Amit Gupta, Senior Director- AI/ML and Mark Wang, Software Engineering PMTS from Salesforce for your contributions in the vision and partnership of cross-Region inference routing on SageMaker AI.
Google Cloud Highlights EMEA Startup Momentum at Slush 2026 | Google Cloud Blog
Published:
As thousands of founders, investors, and technology leaders gather in Helsinki for Slush 2026, the European startup ecosystem is undergoing a profound transformation. The conversation has moved from AI experiments to operating agents.
Google Cloud Highlights EMEA Startup Momentum at Slush 2026 | Google Cloud Blog
Published:
As thousands of founders, investors, and technology leaders gather in Helsinki for Slush 2026, the European startup ecosystem is undergoing a profound transformation. The conversation has moved from AI experiments to operating agents.
Developer tools
3Under the Hood of Native WSL Mode in RubyMine: IJent and EelApi - The JetBrains Blog
Published:
RubyMine’s WSL support has evolved over many years, resulting in several ways of working with projects in WSL. Starting with the 2026.3 release, there is one recommended way to work with a project stored in WSL – Native WSL mode: open or clone the project directly from the Welcome screen and use native WSL interpreters to run Ruby.
Under the Hood of Native WSL Mode in RubyMine: IJent and EelApi - The JetBrains Blog
Published:
RubyMine’s WSL support has evolved over many years, resulting in several ways of working with projects in WSL. Starting with the 2026.3 release, there is one recommended way to work with a project stored in WSL – Native WSL mode: open or clone the project directly from the Welcome screen and use native WSL interpreters to run Ruby.
Vercel Releases UI SDK for Agents to Compose UI from Reusable Behaviors - InfoQ
Published:
Vercel has released UI SDK, an open-source TypeScript library that treats keyboard navigation, focus, selection, popups, and drag and drop as reusable behaviors rather than details buried inside finished components. Vercel UI SDK is published on GitHub as vercel-labs/ui and is currently in beta.
Vercel Releases UI SDK for Agents to Compose UI from Reusable Behaviors - InfoQ
Published:
Vercel has released UI SDK, an open-source TypeScript library that treats keyboard navigation, focus, selection, popups, and drag and drop as reusable behaviors rather than details buried inside finished components. Vercel UI SDK is published on GitHub as vercel-labs/ui and is currently in beta.
Microsoft's official Python agent framework hijacked to distribute malware
Published:
Microsoft’s official agent framework package hosted on PyPI has been hijacked and now installs info-stealing malware, multiple security research firms have discovered.
Microsoft's official Python agent framework hijacked to distribute malware
Published:
Microsoft’s official agent framework package hosted on PyPI has been hijacked and now installs info-stealing malware, multiple security research firms have discovered.
Languages & frameworks
3PostgreSQL: pgjumbo: a new free desktop client for PostgreSQL
Published:
pgjumbo is a new desktop client for PostgreSQL 12+, free to use, for Windows, macOS and Linux, with query, data and administration tools.
PostgreSQL: pgjumbo: a new free desktop client for PostgreSQL
Published:
pgjumbo is a new desktop client for PostgreSQL 12+, free to use, for Windows, macOS and Linux, with query, data and administration tools.
PostgreSQL: pgSherpa: a free PostgreSQL monitoring companion for macOS
Published:
CortexRidge has released pgSherpa, a free desktop companion for PostgreSQL on macOS, now available on the Mac App Store.
PostgreSQL: pgSherpa: a free PostgreSQL monitoring companion for macOS
Published:
CortexRidge has released pgSherpa, a free desktop companion for PostgreSQL on macOS, now available on the Mac App Store.
PostgreSQL: WAL-G 3.0.10 released
Published:
The WAL-G team is pleased to announce the release of WAL-G 3.0.10.
PostgreSQL: WAL-G 3.0.10 released
Published:
The WAL-G team is pleased to announce the release of WAL-G 3.0.10.
Open source
3HomeView 27.2.1 Update Fixes Siri Bugs - MacRumors
Published:
Apple today released a new software update for the HomeView, and it fixes two Siri issues users were running into.
HomeView 27.2.1 Update Fixes Siri Bugs - MacRumors
Published:
Apple today released a new software update for the HomeView, and it fixes two Siri issues users were running into.
Apple Releases Safari Technology Preview 257 With Bug Fixes and Performance Improvements - MacRumors
Published:
Apple today released a new update for Safari Technology Preview, the experimental browser that was first introduced in March 2016. Apple designed Safari Technology Preview to allow users to test features that are planned for future release versions of the Safari browser.
Apple Releases Safari Technology Preview 257 With Bug Fixes and Performance Improvements - MacRumors
Published:
Apple today released a new update for Safari Technology Preview, the experimental browser that was first introduced in March 2016. Apple designed Safari Technology Preview to allow users to test features that are planned for future release versions of the Safari browser.
Apple Releases Siri Remote Firmware Update - MacRumors
Published:
Apple on Tuesday released a firmware update for the fourth-generation Siri Remote. The new version is 0001.0061.0000.0000, up from 0001.0057.0000.0000, according to AppleDB.
Apple Releases Siri Remote Firmware Update - MacRumors
Published:
Apple on Tuesday released a firmware update for the fourth-generation Siri Remote. The new version is 0001.0061.0000.0000, up from 0001.0057.0000.0000, according to AppleDB.
Semiconductors
4Voltage glitch unlocks 80GB on Nvidia's CMP 170HX, but 'faster than an A100' rests on a single benchmark | TechRadar
Published:
Nvidia spent the better part of a decade teaching its silicon to keep secrets from its own owners, but a new piece of hardware research argues that, on at least one chip, those secrets were only ever a power glitch away.
Voltage glitch unlocks 80GB on Nvidia's CMP 170HX, but 'faster than an A100' rests on a single benchmark | TechRadar
Published:
Nvidia spent the better part of a decade teaching its silicon to keep secrets from its own owners, but a new piece of hardware research argues that, on at least one chip, those secrets were only ever a power glitch away.
China threatens 'countermeasures' in the face of new AI GPU rental and export restrictions — Beijing threatens retaliation ahead of Thursday's Xi Jinping meeting, Sino-American trade war may be renewed for a new season | Tom's Hardware
Published:
Yesterday, the Trump administration extended the reach of its existing export controls on high-end AI accelerator chips, namely, but not limited to, Nvidia's Blackwell lineup. The updated regulation now requires licenses for renting those chips in large quantities to the entirety of America's D:5 countries list (nations embargoed from arms sales).
China threatens 'countermeasures' in the face of new AI GPU rental and export restrictions — Beijing threatens retaliation ahead of Thursday's Xi Jinping meeting, Sino-American trade war may be renewed for a new season | Tom's Hardware
Published:
Yesterday, the Trump administration extended the reach of its existing export controls on high-end AI accelerator chips, namely, but not limited to, Nvidia's Blackwell lineup. The updated regulation now requires licenses for renting those chips in large quantities to the entirety of America's D:5 countries list (nations embargoed from arms sales).
US closes the AI chip loophole that let China rent Nvidia hardware abroad - TechSpot
Published:
A Chinese company that can't legally buy a cutting-edge Nvidia processor can still rent time on one, as long as the chip sits in a server rack outside China. The silicon never crosses the Chinese border, but the computing power does.
US closes the AI chip loophole that let China rent Nvidia hardware abroad - TechSpot
Published:
A Chinese company that can't legally buy a cutting-edge Nvidia processor can still rent time on one, as long as the chip sits in a server rack outside China. The silicon never crosses the Chinese border, but the computing power does.
America’s AI Chip War Is Heating Up in the Dumbest Possible Ways
Published:
But under new plans announced on Monday, overseas data centers that want to provide large amounts of computing power to China will have to get special licenses and deal with other annoying new kinds of red tape. The White House also imposed stricter rules on shipments of advanced chips to five countries, including Singapore and Malaysia, where US officials suspect significant levels of smuggling may have taken place.
America’s AI Chip War Is Heating Up in the Dumbest Possible Ways
Published:
But under new plans announced on Monday, overseas data centers that want to provide large amounts of computing power to China will have to get special licenses and deal with other annoying new kinds of red tape. The White House also imposed stricter rules on shipments of advanced chips to five countries, including Singapore and Malaysia, where US officials suspect significant levels of smuggling may have taken place.