Swyx
AI engineering & community
Founder of Latent Space, former Airbyte, Temporal. Writes and podcasts on AI engineering.
Recent activity
-
Arca, an AI-native wealth management firm, uses agents to perform long-horizon financial work, such as analyzing client data and generating financial plans that can take up to ten days to complete. The firm has developed an internal memory system to track the progress of these tasks and ensure that human advisors are involved in the decision-making process. Arca's agents are designed to prompt humans for input, rather than the other way around, and the firm has implemented various safety measures to prevent agents from making decisions that could harm clients. AI summary
Read more → -
Google's AI Mode has reportedly led to a -15% drop in search revenue, equivalent to a $500B loss, the worst since its 2004 IPO, with the AI Bear Market potentially looming. AI summary
Read more → -
The AIE CODE conference, the top Coding Agent conference, features three themes: Scaling without Slop, System One models, and Increasing Ambition, with a focus on showcasing the latest advancements in agentic coding and cybersecurity. AI summary
Read more → -
Elorian's Glimpse model has demonstrated 50.48% accuracy on the DecoyCount benchmark, surpassing frontier models like GPT-6 Astra (43.17%) and Gemini-3.8-flash (32.71%). This achievement highlights Glimpse's ability to visually reason and ground its counts in visible evidence, rather than relying on shortcuts or substitution heuristics. The DecoyCount benchmark, co-released by Elorian, tests whether language models genuinely count objects or rely on visual shortcuts, revealing a gap between seeing and reasoning. AI summary
Read more → -
Google's Gemini 4 Carbon, a new frontier model, has topped leaderboards in cybersecurity and long-horizon tasks, with benchmarks including 88.2% on SEC-bench Pro for vulnerability finding and 71.0% on CWE-bench for patching, with access initially limited to trusted cyber defenders. AI summary
Read more → -
Supabase's agent-friendly features, such as local backend setup without Docker, declarative schemas, and compute alongside the database, simplify the development process for AI models and reduce friction. These improvements make it easier for developers to understand and work with the backend, aligning with the needs of agents that repeatedly create and test environments. By providing a more straightforward backend experience, Supabase enables agents to work more efficiently and effectively. AI summary
Read more → -
Anthropic has ended coding interviews for job applicants, citing that hunger for innovation beats tenure in hiring. This shift aims to prioritize candidates with impactful ideas and business value over technical skills demonstrated through coding challenges. AI summary
Read more → -
Anthropic has developed a modular architecture for their Claude platform, separating the "brain" (managed agents) from the "hands" (API), allowing for more efficient and secure model deployment. This separation enables better isolation, credentials management, and network controls, and also facilitates lower latency and better reliability. AI summary
Read more → -
Katelyn Lesse and Angela Jiang join Latent Space to explain the engineering behind the Claude Platform: launching Opus 5.5, separating the model’s “brain” from the agent’s “hands,” scaling APIs, and building reliable managed agents. A new introduction ...
Read more → -
Developers can set up an offline coding agent using OpenCode, Ollama, and Qwen by downloading the necessary models and dependencies, setting up the project configuration, and running the coding tasks locally without internet connectivity. This setup allows for tasks such as code review, debugging, and testing to be performed offline, making it suitable for remote work or situations where internet access is limited. AI summary
Read more → -
Anthropic has filed for a $2 trillion IPO with a 1400% growth rate, generating 48% of revenue through cloud partners, and boasting an 85% margin, with $1 billion in operating profit in Q2 and $134 billion in cash reserves. AI summary
Read more → -
Kimi 3.1 Pro 104BA-2.8T, a post-trained version of Kimi K3, has achieved a Fable-5.1-class open-weight model, narrowing the Closed-AI lead to 1 month, with a 2-3x speed advantage in the Pro Highspeed API. AI summary
Read more → -
AI models are solving more mathematical problems. A pattern is emerging in their successes that could help us predict which problems they’ll solve next.
Read more → -
The "Black October" incident wave involved a cluster of critical disclosures and breaches over 48 hours, including PHP2Shell, a PHP unauthenticated RCE, and a CVSS 10 flaw in GitLab's releases since 2014. Several AI models and tools were released, including Kitten-TTS 2, a 1.58-bit open-source TTS model, and Claude, which achieved formal verification at scale, proving security properties across over 100 open-source libraries in weeks. AI summary
Read more → -
Modal's CEO Erik Bernhardsson explains that traditional GPU compute contracts are "rectangular" and don't account for the dynamic nature of AI workloads, which can grow, spike, and fall throughout the day. This leads to wasted resources and idle hardware. Modal's approach to providing a cloud-based platform for AI infrastructure solves this problem by pooling demand across thousands of customers, securing capacity underneath, and allowing each customer to use what they need when they need it. AI summary
Read more → -
Modal's CEO Erik Bernhardsson explains that traditional GPU contracts are rectangular in nature, where a customer pays for a fixed number of GPUs for a fixed period. In contrast, real workloads grow, spike, and fall throughout the day, making it difficult for customers to predict and manage their GPU usage. Modal addresses this by pooling demand across thousands of customers and securing capacity underneath, allowing each customer to use what they need when they need it. AI summary
Read more → -
Anekha Sokhal, CEO of Moshi, discusses how her background in mathematics and JPMorgan's energy derivatives trading informed her approach to building a recommendation engine for financial traders. Moshi processes vast amounts of data, including structured and unstructured data, to provide real-time insights and recommendations, similar to a personalized dashboard, and aims to solve the problem of how to make data actionable for traders. AI summary
Read more → -
Anekha Sokhal, CEO of Moshi, is building a recommendation engine that connects market information to what matters for a trader's book, processing more data than humans can read in 4 seconds. Moshi's system analyzes structured and unstructured data, including time series, technical analysis, and news, to provide actionable insights to traders. AI summary
Read more → -
OpenAI's Patch the Planet program, in collaboration with another party, discovered PHP2Shell vulnerabilities in core PHP affecting 70% of the web, using GPT-5.6 Sol, and coordinated a response over two months. Researchers identified two PHP-engine memory-corruption vulnerabilities with demonstrated unauthenticated exploitation against major CMSs. AI summary
Read more → -
Researchers at OpenAI's Patch the Planet program, in collaboration with Wiz, discovered PHP2Shell vulnerabilities in core PHP affecting 70% of the web using GPT-5.6 Sol. The vulnerabilities were found after approximately two months of coordinated response, and a patch was released, with Cloudflare deploying initial WAF mitigations before the patch release. AI summary
Read more →