Mark Zuckerberg told Meta employees that AI agents haven't progressed as quickly as he'd hoped, but a closer look at the data suggests the real problem is how agents are tested. Scale AI found frontie…
Meta's Muse Spark scored 42.8 on HealthBench Hard, more than double Gemini 3.1 Pro's 20.6, while remaining free on meta.ai and pricing its API at $1.25 per million input tokens and $4.25 per million o…
A mid-2026 audit of top-down AI-assisted software engineering shows agents absorbing implementation work on schedule, with SWE-bench Verified scores rising from 1.96% in October 2023 to near saturatio…
A developer built Humans vs. HLE, a web-based quiz that pits humans against frontier AI models on Humanity's Last Exam, a benchmark of expert-level academic questions. The app runs on Cloudflare Worke…
Boris Cherny, creator of Anthropic's Claude Code, outlined a four-step framework for enterprise AI adoption on X, arguing that companies should measure AI returns by comparing engineering effort saved…
Meta released Muse Spark 1.1 on OpenRouter on July 9, but restricted access to US developers only, according to Meta's developer announcement. The open-weight, multimodal coding model is priced at $1.…
Amazon is closing Mechanical Turk to new users on July 30, putting the 21-year-old platform into maintenance mode after a 2023 study found that 33–46% of workers were using large language models to co…
A friend of the author failed a senior AI engineer interview at an unnamed company after being asked what happens when a retriever returns a document contradicting the user's intent, a question that n…
Instagram CEO Adam Mosseri said he looks for three core traits in employees—quick learners, gritty, and self-aware—during a recent episode of 'Lenny's Podcast.' Mosseri emphasized that these traits ma…
A July 2026 analysis of 105 finance job postings reveals a gap between CFOs' stated preference for embedded AI from existing vendors and actual hiring for custom-built AI agents, with 70% of corporate…
Finance systems roles are transforming into engineering jobs, with companies like Faire, Scale AI, and Pigment hiring for positions that require building AI agents and automated workflows for tasks su…
OpenAI retracted its recommendation for Scale AI's SWE-Bench Pro coding benchmark after an audit found roughly 30% of its 731 tasks were broken due to overly strict tests, underspecified prompts, low-…
Mercor, a three-year-old AI training marketplace run by a 23-year-old CEO, is in talks to raise at a $20 billion valuation, doubling its worth in nine months despite a recent data breach that cost it …
OpenAI retracted its recommendation of SWE-Bench Pro after an audit found roughly 30% of its 731 tasks are broken. The audit, combining AI investigator agents and human reviewers, identified 200–249 b…
Former OpenAI researcher Phil Chen advises professionals to focus on problem selection, time, relationships, and reputation to thrive in the AI era, arguing that human intuition in choosing impactful …
SemiAnalysis projects Meta will hold more AI compute capacity than OpenAI or Anthropic by end of 2026, driven by multi-gigawatt datacenters and a $14.3 billion Scale AI stake. The firm also predicts M…
Meta announced Muse Spark 1.1, a multimodal AI model designed for agentic tasks, with improvements in computer use, coding, and multimodal understanding. The model, released by Meta's Superintelligenc…
Meta's AI division, Meta Superintelligence (MSL), has made significant progress in the past year following a rebuild after the Llama 4 failure, with advantages in data, talent, and compute that positi…
Wildcard, an agentic commerce optimization platform for e-commerce and retail brands, is hiring its first founding engineer. The company, which grew 50% month over month, helps brands manage how their…
Meta released Muse Spark 1.1, claiming it surpasses prior AI models from OpenAI, Anthropic, and Google on some benchmarks, though it still lags behind flagship models. The update comes under new chief…