programming

According to NetEase's report on "DeepTech," the open-source AI programming tool OpenCode has officially confirmed it was breached, affecting approximately 1.4% of OpenCode Console users, with some API keys for models stored on the platform potentially leaked. Multiple developers reported unusual charges on accounts used to call models from OpenAI, Anthropic, and others, with some claiming losses of hundreds of dollars and others stating amounts exceeding $1,700. The officials stated that the incident is related to a large-scale distillation attack and are currently working with OpenAI, Anthropic, and Google to address the issue.OpenCode was developed by members of the Anomaly team, with one of the main creators being Dax Raad. The team has previously worked on projects like SST and OpenNext. In addition to local programs, its cloud service Console can centralize API keys from providers like OpenAI, Anthropic, and Google in a workspace and call them on behalf of users; the incident occurred on this cloud service. A notification email stated that the attacker gained unauthorized access to a set of decommissioned old systems, and some third-party model API keys in the Console workspaces may have been exposed. Developer Colin Parsons reported unusual charges of about $281 on his OpenAI account and about $168 on his Anthropic account, while Preston Zeller reported losses of around $1,700. Other users reported unusual charges of about $400 on Anthropic and account restrictions triggered by unusual usage on Google services.OpenCode has notified affected users to revoke and replace their keys and check recent call records. Lydia Hallie, a member of the Anthropic Claude Code team, stated that Anthropic has collaborated with OpenCode and revoked the Claude API keys believed to be affected. Tibo, a member of the OpenAI team, stated that they are working with OpenCode to help affected API users restore access and refund the amounts that were unusually consumed during the attack. Dax Raad clarified that the affected service is OpenCode Console, not the locally running OpenCode, and that keys configured directly in the local tool will not be sent to OpenCode servers. On October 18, he stated on social media that the attacker had been continuously attempting to breach the system for the past 48 hours, and the team is strengthening defenses.

According to CoinDesk, Cardano founder Charles Hoskinson stated that he is personally developing a new programming language for the Cardano network and collaborating with Input Output Global engineer Lucas Rosa. Lucas Rosa is a co-creator of Aiken, one of Cardano's main smart contract languages. Hoskinson posted on the X platform that this will be the first programming language to combine user transaction completion intentions with cryptographic proofs.It is reported that the language will support three tools: transaction intent, data with carry proofs, and Groth16 proofs. Transaction intent allows users to specify the goals they want to achieve, while leaving the specific steps to the application, for example, traders can request that a token be exchanged for another only when the final price meets certain conditions, without specifying the exact path.Data with carry proofs can attach cryptographic evidence to information, proving that it complies with specific rules, allowing other applications to verify without recalculating. Groth16 is a widely used proof system that can generate small-sized, fast-verifying proofs, demonstrating that conditions have been met without exposing private data.With the combination of the three, developers can build applications that coordinate transactions, check complex conditions, and pass verified information between services, ultimately verified by Cardano with a small proof. Currently, Cardano developers mainly use Haskell-based Plutus and Aiken to write smart contracts, while Input Output Global, founded by Hoskinson, has its own privacy blockchain Midnight with its language Compact. The name, public code, and release date of this new language have not yet been announced.

OpenCode announces a limited-time free access to DeepSeek V4.1 Flash. Users can select DeepSeek V4.1 Flash Free from the model list, with no charges for input and output tokens. The free version has a context limit of 200,000 tokens, while the paid version has a limit of 1,000,000 tokens, which is an 80% reduction, making it easier to encounter context limits when processing large code repositories or running a Coding Agent for extended periods.DeepSeek V4.1 Flash was integrated into OpenCode on September 10. On October 12, OpenCode will increase the usage quota for V4.1 Flash in the $10 Go package to a value of $80, and subsequently add an independent free tier. The official documentation confirms that this free access has a time limit, but the end date has not yet been announced, and states that the free tier service provider adheres to a zero data retention policy and will not use user content for model training.

When discussing the reasons for acquiring the AI programming company Cursor, Musk stated that to create truly outstanding superintelligence, a large amount of training data is required, and programming must be conquered.Musk mentioned that, to his knowledge, there are only three organizations with sufficient programming training data: OpenAI, Anthropic, and Cursor, with Google barely making the list. Musk's SpaceX completed the acquisition of Cursor on August 14, with a transaction price of approximately 60 billion dollars.

Anthropic released Claude Sonnet 5.5 this Monday, as an upgraded version of Sonnet 5 launched in June this year. The official statement claims that this mid-range model operates over 30% faster than the previous generation, excelling in clearly defined daily tasks, fixing defects, and creating documents, slides, and spreadsheets, while also possessing sharp design capabilities. Its pricing remains at $2 per million input tokens and $10 per million output tokens, which is half of Opus 5.5, and the token consumption per task is nearly one-third less than Sonnet 5.In the Terminal-Bench 4 test, Sonnet 5.5 scored 70.6%, higher than Opus 5.5's 66.4%, and Sonnet 5's score was only 10.3%. Results from the independent testing organization Artificial Analysis also show its lead, recording 63.6% and 59.6%, while OpenAI's GPT-6 Astra scored 59.1%. In the GDPval-AA assessment of real job performance across 44 occupational categories, Sonnet 5.5 received a score of 1844, which is roughly on par with Opus 5.5's score of 1846, while GPT-6 Sol scored 1487.

Analysis of the AI semiconductor and infrastructure cycle indicates that the demand for AI infrastructure will continue to exceed that of the internet era, as user penetration and per capita token are multiplied and converted into tokens, significantly raising the ceiling. User penetration has surpassed 50%, with growth primarily coming from the still-early per capita token; the median monthly AI spending per employee in U.S. companies is about $12, which could long-term approach around 10% of white-collar salaries, approximately $1,000 per month, leaving nearly two orders of magnitude of space in between. Unlike the flat subscriptions and extremely low marginal hardware consumption of the internet, the high costs of inference make the marginal cost of a single access higher, requiring greater infrastructure intensity.After developers program, the next ARR growth will still mainly come from broad programming: non-programmers use programming infrastructure to complete non-programming tasks across industries, with programming becoming the default execution kernel for agents. Tasks related to broad programming account for about 60% to 70% of ARR. As of June 2026, Codex accounted for 64% of the total output tokens from Codex and ChatGPT among OpenAI's enterprise clients; since February, Codex has seen a much higher weekly growth in verticals such as law, sales recruitment, and marketing compared to engineering. In Anthropic's revenue, narrow development/software accounts for about 40%, while finance and insurance exceed 20%, with law, life sciences, retail, and others also having considerable shares.The demand-side token growth logic remains, with a high overlap between funders and beneficiaries.

Theo Hourmouzis, General Manager of Anthropic for Australia and New Zealand, stated at an event hosted by the National Australia Bank that the company's Australian clients lag behind their global counterparts in utilizing the programming capabilities of the Claude chatbot, but are more inclined to use it for simple tasks.Hourmouzis cited Anthropic data indicating that approximately 5.8% of Claude's usage in Australia pertains to basic tasks such as invoice processing, inventory management, and general administration, which is higher than the global level of about 4.7%. He pointed out that local usage tends to lean more towards backend and practical work.

Swedish Stockholm AI programming startup Lovable announced the completion of a $400 million Series C funding round, bringing its post-funding valuation to $13.3 billion. This round of financing was co-led by Menlo Ventures and the EU-backed investment institution Scaleup Europe Fund (managed by EQT AB), with several venture capital firms participating, as well as new investors from Latin America and Asia, including Tencent Holdings.Lovable is dedicated to the field of AI-assisted programming and is accelerating its expansion to compete with larger rivals such as Anthropic.

AMD officially launched the on-premises AI programming platform AMD Instinct Coder for enterprises, integrating AMD EPYC processors, AMD Instinct GPUs, Supermicro AI servers, Spectro Cloud PaletteAI inference launch platform software, and the GLM-5.2 model optimized by AMD inference microservices, forming a comprehensive solution that is pre-integrated from hardware to software and can be quickly deployed.AMD stated that the total cost of ownership of this platform can reduce costs by up to 70% compared to using cloud-based cutting-edge models, with a capital recovery period as short as 6 months, directly addressing the two core pain points that enterprises commonly encounter in AI programming assistant pilots: the cost impact of cloud models and the data security risks of third-party services.Spectro Cloud's PaletteAI platform provides intelligent routing capabilities, allowing for strategic automatic switching between on-premises inference and cloud-based cutting-edge models from Anthropic, OpenAI, Google, and xAI, and is equipped with token quotas, audit tracking, and Grafana dashboards to visualize cost comparisons between on-premises and cloud environments, fully compatible with standard operating systems and Kubernetes. Each node can support up to 50 users and is compatible with mainstream development tools such as Claude Code, OpenAI Codex, Visual Studio Code, and Cursor.IDC executives commented that this platform "directly addresses the core needs of enterprises wanting AI programming productivity without the unpredictable cloud costs or risks of exposing source code." AMD's move shifts the AI programming assistant from a pure service model billed by usage to an infrastructure investment model with pre-planned costs.

Vitalik posted on the X platform stating that a new type of "high-level programming language" worth trying is a language compiled to Lean (or HOL, etc.), focusing on making it as easy as possible for humans to read definitions and theorems. Rather than proofs, because proofs only need to be correct, the key lies in the definitions and theorems themselves. The envisioned use is that AI outputs a large segment of proof, and the reader needs to understand as easily as possible which precise claims have actually been proven in these outputs.

The international authoritative consulting agency IDC released the "2025 China AI Programming Market Share Report." The report shows that the scale of China's AI programming market will reach 399 million yuan in 2025, and it is expected to grow rapidly to 1.173 billion yuan by the end of 2026.In terms of market competition, Alibaba ranks first with a market share of 47.6% thanks to its core AI programming product Qoder, a figure that exceeds the combined market shares of the second to fifth ranked companies.IDC analysts pointed out that AI Coding is evolving from a mere "code generation tool" into a "software productivity platform" that covers the entire software lifecycle. Currently, Alibaba's Qoder product line has served hundreds of thousands of enterprises including China FAW, CITIC Securities, and AsiaInfo Technologies, with over 5 million users globally.

According to TechCrunch, Indian AI programming startup Emergent has completed a $130 million Series C funding round, with a post-money valuation of approximately $1.5 billion, an increase of about 5 times from a $300 million valuation in January this year. This round was led by private equity firm Creaegis, with participation from MNI Ventures-Claypond, Sentinel Global, as well as Khosla Ventures, SoftBank Vision Fund II, Lightspeed, Y Combinator, among others, bringing the total funding to $230 million.Emergent focuses on providing an "engineering team as a service" AI programming platform for small and medium-sized enterprises and entrepreneurs. It currently has an annual revenue of approximately $120 million, with a growth of about 70% in the past four months, and over 200,000 paying users, serving industries including logistics, manufacturing, construction, and property management.

According to a recent article by npm co-founder Laurie Voss, research from Stanford University's Digital Economy Lab based on ADP payroll data shows that since the end of 2022, the number of employed junior software developers aged 22 to 25 has decreased by 19%, and entry-level software development positions have dropped by 28% from their peak, with the unemployment rate for computer science graduates rising to 6.1%. However, the total number of developers employed in the U.S. has still grown by 4.4% during the same period, with the employment of senior developers aged 41 to 49 increasing by 14%. Data from the U.S. Bureau of Labor Statistics (BLS) further indicates that over the past year, the number of "computer programmer" positions primarily responsible for writing code on demand has decreased by 16%, while positions for data scientists and core software developers that require more architectural judgment have increased by 12% and 2%, respectively.At the same time, the proliferation of AI tools has led to an explosion in software creation. GitHub added a record 36 million accounts and 121 million code repositories last year, and the number of app submissions to the Apple App Store surged by 80% year-on-year in the first quarter of 2026. Data from platforms like Vercel and Lovable indicates that over 60% of new users are "non-traditional developers" such as product managers and analysts. Industry analysts warn that as AI replaces basic coding tasks, the traditional "junior to senior" engineer apprenticeship promotion path has been disrupted, raising concerns about the safety of AI-generated code and challenging the future sources of senior developers. However, the latest hiring data from platforms like Indeed shows that the demand for related entry-level positions hit bottom in May 2025 and has begun to show signs of rebound.

Visa has invested in the AI programming platform Replit to promote the development of agentic payments and explore how developers and AI agents can directly receive payments within the platform.

According to Bloomberg, AI programming startup Cognition AI announced the completion of a new funding round of over $1 billion, with a post-funding valuation of $26 billion, more than doubling the valuation from the last funding round in September 2025.This round of financing was co-led by Lux Capital, General Catalyst, and 8VC, with follow-on investments from Ribbit Capital, Atreides Management LP, and Peter Thiel's Founders Fund, among others.

According to foreign media citing informed sources, Cognition AI is in the early negotiation stage of a new round of financing. The company is leveraging the growing demand for software development generative artificial intelligence technology, and this round of financing will more than double its valuation to reach $25 billion.One informed source stated that the goal of this AI programming startup is to raise hundreds of millions of dollars or more. Negotiations are ongoing, and the terms may change.

SpaceX posted on the X platform that it has reached a cooperation intention with AI programming startup Cursor, which grants SpaceX the right to acquire it later this year for $60 billion, or to pay $10 billion to advance cooperation between the two parties. Cursor is a rapidly growing artificial intelligence startup focused on developing code writing software.