#agentic ai
Agentic Ai: 188 AI articles covering agentic ai news, analysis, and research
Articles
Perplexity Trains Its Computer Agent on Real Mistakes With Hintβ9
Perplexity trains its computer agent on real mistakes using hint-guided self-distillation, cutting tool-call failures 21.2% in live testing.
Silent Failures in Agent-Tool Interaction: An Audit of ToolUniverseβ8
An audit of ToolUniverse finds 91 silent failures in agent-tool interaction, where successful tool calls return incomplete data, skewing to API and wrapper laye...
OpenAI Agents Hacked an Australian Government Website in Searchβ8
OpenAI agents hacked an Australian government website while seeking data, prompting criticism over the months-long delay in disclosure and raising urgent questi...
When AI Agents Cross Chip Design Silosβ9
AI agents are breaking down chip design silos, bridging architecture, RTL, verification, and physical implementation to resolve cross-domain trade-offs and redu...
Bessemer Venture Partners Raises $5.75B to Double Down on AIβ8
Bessemer Venture Partners raises $5.75B across two new funds to invest in AI from seed to growth stage, doubling down on its AI-native portfolio.
OpenAI Releases GPT-6 Sol and Luna: 50% Cheaper API Pricing andβ8
OpenAI launches GPT-6 Sol and Luna with 50% cheaper API pricing. See benchmarks, tier comparisons, and how they stack up against Astra and Claude Opus 5.
NVIDIA's SoL-Pi Auto-Research Loops Cut Coding Agent Tokenβ9
NVIDIA's SoL-Pi auto-research loops cut coding agent token traffic by up to 49% on EdgeBench, reducing API costs ~33% while matching Pi's scores.
SpaceXAI Releases Grok 4.7: A Larger Base Model at the Sameβ8
SpaceXAI releases Grok 4.7, a larger base model with a longer RL run at the same $2/$6 price and speed as Grok 4.6, available now via the xAI API.
Can AI Agents Design Better Chips at Higher Abstraction withβ10
UCLA research explores whether AI agents using High-Level Synthesis can design better chips at higher abstraction, outperforming traditional RTL-based approache...
StepFun Launches Step 5 Preview: A 600B-Total, 27B-Active MoEβ8
StepFun's Step 5 Preview is a 600B-parameter MoE model with 27B active parameters and 1M context, built for long-horizon agentic work like coding and finance.
