惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

The GitHub Blog
The GitHub Blog
S
SegmentFault 最新的问题
L
LangChain Blog
Simon Willison's Weblog
Simon Willison's Weblog
N
News and Events Feed by Topic
Exploit-DB.com RSS Feed
Exploit-DB.com RSS Feed
WordPress大学
WordPress大学
M
MIT News - Artificial intelligence
A
About on SuperTechFans
Microsoft Security Blog
Microsoft Security Blog
I
Intezer
Know Your Adversary
Know Your Adversary
H
Heimdal Security Blog
博客园 - 叶小钗
B
Blog RSS Feed
F
Fortinet All Blogs
Hacker News: Ask HN
Hacker News: Ask HN
A
Arctic Wolf
小众软件
小众软件
Help Net Security
Help Net Security
MongoDB | Blog
MongoDB | Blog
aimingoo的专栏
aimingoo的专栏
G
Google Developers Blog
Forbes - Security
Forbes - Security
Latest news
Latest news
AI
AI
I
InfoQ
H
Hackread – Cybersecurity News, Data Breaches, AI and More
C
CXSECURITY Database RSS Feed - CXSecurity.com
W
WeLiveSecurity
C
Cybersecurity and Infrastructure Security Agency CISA
人人都是产品经理
人人都是产品经理
Cyberwarzone
Cyberwarzone
奇客Solidot–传递最新科技情报
奇客Solidot–传递最新科技情报
J
Java Code Geeks
Engineering at Meta
Engineering at Meta
C
Cyber Attacks, Cyber Crime and Cyber Security
O
OpenAI News
博客园 - 【当耐特】
T
Threat Research - Cisco Blogs
GbyAI
GbyAI
U
Unit 42
D
Darknet – Hacking Tools, Hacker News & Cyber Security
G
GRAHAM CLULEY
Apple Machine Learning Research
Apple Machine Learning Research
宝玉的分享
宝玉的分享
Google DeepMind News
Google DeepMind News
T
Threatpost
T
The Blog of Author Tim Ferriss
罗磊的独立博客

Swift for Visual Studio Code comes to Open VSX Registry | InfoWorld

Notion courts developers with a platform for AI agents and workflow automation Using continuous purple teaming to protect fast-paced enterprise environments A better way to work with SQL Server Evidence-driven workflows: Rethinking enterprise process design AWS debuts Graviton-powered Redshift RG instances to cut analytics costs SAP’s AI promises last year? Most are still rolling out First look: Lemonade serves up local AI with limitations GitLab CEO sees developer tool bill increasing 100-fold Red Hat adds support for agentic AI development What’s new and exciting in JDK 26 Kill the loading spinner with local-first data and reactive SQL A networking revolution at AWS Tokenmaxxing is super dumb How to add AI to an existing product (without annoying users) Your AI doesn’t need another database What happens when engineering teams reorganize around AI agents Python isn’t always easy When cloud giants meddle in markets 12 model-level deep cuts to slash AI training costs The best new features in Python 3.15 Teradata launches platform for enterprise AI agents moving beyond pilots Three skills that matter when AI handles the coding MongoDB targets AI’s retrieval problem Building AI apps and agents with Microsoft Foundry Designing front-end systems for cloud failure No, AI won’t destroy software development jobs Diskless databases: What happens when storage isn’t the bottleneck Vibe coding or spec-driven development? The agentic AI distraction Vibe coding or spec-driven development? How to choose Cloud providers are blinded by agentic AI SAP to acquire data lakehouse vendor Dremio Small language models: Rethinking enterprise AI architecture Making AI work through eval hygiene Improving AI agents through better evaluations AI in the cloud is easy but expensive Running AI in the cloud is easy – and expensive Making AI work for databases Harness teams of agentic coders with Squad Harness teams of coding agents with Squad Oracle NetSuite announces AI coding skills for SuiteCloud developers Why it’s so hard to create stand-alone Python apps A new challenge for software product managers The hidden cost of front-end complexity GitHub shifts Copilot to usage-based billing, signaling a new cost model for enterprise AI tools OpenAI’s Symphony spec pushes coding agents from prompts to orchestration The front-end architecture trilemma: Reactivity vs. hypermedia vs. local-first apps Enterprise AI is missing the business core The best JavaScript certifications for getting hired Google begins putting the guardrails on agentic AI Why world models are AI’s next frontier Where to begin a cloud career Google pitches Agentic Data Cloud to help enterprises turn data into context for AI agents Is your Node.js project really secure? How I doubled my GPU efficiency without buying a single new card SpaceX secures option to acquire AI coding startup Cursor for $60B Google’s Gemma 4 shines on local systems – both big and small AI is upending the SaaS game How AI is upending SaaS tools Snowflake offers help to users and builders of AI agents From the engine room to the bridge: What the modern leadership shift means for architects like me Addressing the challenges of unstructured data governance for AI The cookbook for safe, powerful agents Enterprises are rethinking Kubernetes GitHub pauses new Copilot sign-ups as agentic AI strains infrastructure Best practices for building agentic systems Making agents dull Oracle delivers semantic search without LLMs When cloud giants neglect resilience Exciting Python features are on the way Ease into Azure Kubernetes Application Network The agent tier: Rethinking runtime architecture for context-driven enterprise workflows The two-pass compiler is back – this time, it’s fixing AI code generation MuleSoft Agent Fabric adds new ways to keep AI agents in line Salesforce launches Headless 360 to support agent‑first enterprise workflows Tap into the AI APIs of Google Chrome and Microsoft Edge Where will developer wisdom come from? GitHub adds Stacked PRs to speed complex code reviews The hyperscalers are pricing themselves out of AI workloads HTMX 4.0: Hypermedia finds a new gear Google Cloud introduces QueryData to help AI agents create reliable database queries Hands-on with the Google Agent Development Kit Are AI certifications worth the investment? AWS targets AI agent sprawl with new Bedrock Agent Registry Cloud degrees are moving online Swift for Visual Studio Code comes to Open VSX Registry AI agents aren't failing. The coordination layer is failing How Agile practices ensure quality in GenAI-assisted development Anthropic rolls out Claude Managed Agents Microsoft’s reauthentication snafu cuts off developers globally Meta’s Muse Spark: a smaller, faster AI model for broad app deployment Bringing databases and Kubernetes together Rethinking Angular forms: A state-first perspective Minimus Welcomes Yael Nardi as CBO to Facilitate Strategic Growth Microsoft announces end of support for ASP.NET Core 2.3 Get started with Python’s new frozendict type AWS turns its S3 storage service into a file system for AI agents Microsoft’s new Agent Governance Toolkit targets top OWASP risks for AI agents The winners and losers of AI coding GitHub Copilot CLI adds Rubber Duck review agent
How open source ideals must expand for AI
2026-04-23 · via Swift for Visual Studio Code comes to Open VSX Registry | InfoWorld

Open source has never been just a licensing model. Rather, it’s also a philosophy about shared effort, shared transparency, and shared agency. The shared goal is to make an impact in the world. In the age of AI‑assisted development and agents, there is a line of thinking that AI slop, specifically mass-produced and submitted code, is the downfall of open source projects. On the contrary, I think open source is headed for a resurgence like we’ve never seen before, as long as we emphasize the aspects of open source beyond code submissions.

In this new world, the philosophy, ethics, and morals of open source are more relevant than ever. However, the focus of open source needs to evolve past raw code: Specification files (spec files) and governance documents (constitutions) are becoming as important as the source itself. The challenge is not to choose between open source and AI, but to recognize that open source is now a community-based control and scope mechanism for open technologies.

Let me break this down further. Specs describe intent and outcomes, while code shows how that intent is actually realized. When something goes wrong, you still need to trace the path from spec to implementation. Trust is earned, not inferred, so the promise that, for example, an app values your privacy or that an agent never sends data to third parties must be backed by code and built on pipelines that anyone can inspect and verify (e.g. Acquacotta Constitution).

Governance complements spec files, in showing how a spec is created, enforced and followed. It’s the “people decisions” around a project — who makes the final decision? If there’s a vote on something, who votes? How do they vote? It’s these seemingly pedantic yet crucial decisions that emerge from governance, and they are the backbone for how spec files are created and followed when code contributions become as simple as writing a basic agent prompt.

Open means open, even for AI

The main criticism of AI with open source is that code contributions become open to everyone, not just those with deep technical knowledge. This makes the pillars of an open source community outside of the code more important than ever. Users fundamentally become contributors when anyone can create code, which means they need agency in specification. Additionally, this new class of community members needs to be able to help influence and change governance and spec, just like “normal” contributors in the pre-AI days. The spec files submitted with their AI-generated code must be open for inspection and reproducible, just like a more traditional code contribution. The ability to fork the implementation and run it on outside infrastructure also remains, enabling contributors to further refine their own customizations, integrations, and optimizations.

This is how organizations retain real agency in an era where code is a commodity. In other words, spec files broaden what we need to keep open; they do not replace the need to keep code, build systems, and dependency trees open and inspectable. The future is not specs instead of open source; it is open source plus open specs and open governance.

You version, review, and discuss specs in the same way you review and discuss code. You make architecture and governance artifacts part of the public record so that others can learn from, reuse, and improve them. This creates a richer set of open source assets. The repository does not just contain code; it contains the constitution of the project, the architectural reasoning, and the guardrails that keep AI tools and AI-driven code contributions on the rails.

This lowered barrier cannot be overstated. Truly anyone can now contribute, not just those who understand the code-based components of a project. Domain experts, designers, and operations specialists can propose changes at the spec level that AI agents then help implement. For a community that leads with open source values, and robust testing frameworks (coded in the constitution), this fuels the creation of high‑quality software while preserving the transparency, reviewability, and forkability that open source depends on.

‘Real’ open source

We do risk the convergence of open source and specs turning into a two-sided purity test. Some argue that if AI wrote most of the code, it is not “real” open source. This argument implies that the provenance of each line of syntax matters more than whether the system as a whole is transparent, forkable, and governed in the open. Meanwhile, the opposing side suggests that if we can regenerate the implementation from a spec, we no longer need to worry as much about licensing and code openness — as if an elegant constitution makes inspection of the actual machinery optional.

Both positions miss the point. If you care about user agency, security, and long‑term sustainability, as all open source projects should, you need both open code and open build pipelines, so anyone can inspect, reproduce, and harden what is running. You need open specs and governance, so anyone can understand what the system is supposed to do, how it is supposed to behave, and how decisions get made over time.

The new “definition” of open must consider implementation, specification, and governance as three critical factors that must be woven together. Open implementation means the source, dependencies, and build system are available under an open source license so you can rebuild, audit, and run the software yourself. Open specification means the requirements, architecture, and project constitution are documented, versioned, and public, so others can reuse them, learn from them, and adapt them to their own needs. Open governance means the processes by which changes are proposed, reviewed, and accepted — whether at the spec level or in code — are transparent and participatory.

The path forward for open source communities is not to retreat from spec‑driven, AI‑assisted development, nor to declare the old mission obsolete. It is to lead in defining and practicing what open specification, governance, and implementation look like together in an AI‑first world — and to do so with the confidence to dream bigger than incremental automation.

It’s this ability for individuals and organizations to dream bigger that makes it possible to tackle problems that were previously too big, too weird, or too niche to justify a full project team. A UX designer who’s never committed a line of code can suddenly create a serious tool. A security engineer can prototype a new threat‑hunting pipeline end‑to‑end. An architect can stand up a reference implementation of a safer integration pattern, then let others clone and extend it. In each case, AI doesn’t replace expertise; it amplifies it. The most valuable people in the room spend less time grinding out boilerplate and more time on intent, constraints, and trade‑offs.

AI tooling isn’t a one way street

However, there’s one really important thing to remember about dreaming bigger with AI tools: You are not the only one who has easy access to them — so do your competitors. Perhaps more importantly, the bad guys have them too.

When it comes to the former, dreaming bigger means dreaming bigger than the folks who would take business away from you. It’s about making sure that your customers stay your customers and that you can attract new ones on a continual basis.

When it comes to the latter, it means considering that the people who would do you and your customers harm have these same tools. And, if anyone dreams bigger, it’s the people who want to break your systems, defraud your customers, and manipulate your users.

That’s one reason the traditional “open source = open code” framing is starting to feel too small. If attackers can continuously remix powerful AI tooling in the shadows, defenders need open, inspectable patterns for detection, response, and governance that anyone can adopt and improve.

Open source’s big tent

The same openness that once turned a loose collection of hackers into the engine of the modern software stack can now be applied to a new layer of specs and agents. This is a new way forward to make open source even more accessible to a new type of contributor. For open source, it’s not about fighting AI slop. Instead, we as a global community need to push the positives of AI-driven contributions forward, so that the good far outweighs the bad.

New Tech Forum provides a venue for technology leaders—including vendors and other outside contributors—to explore and discuss emerging enterprise technology in unprecedented depth and breadth. The selection is subjective, based on our pick of the technologies we believe to be important and of greatest interest to InfoWorld readers. InfoWorld does not accept marketing collateral for publication and reserves the right to edit all contributed content. Send all inquiries to doug_dineley@foundryco.com.