Thariq Shihipar
PulseAugur coverage of Thariq Shihipar — every cluster mentioning Thariq Shihipar across labs, papers, and developer communities, ranked by signal.
6 day(s) with sentiment data
-
Discovered Materials raises $9M for AI-driven chip material discovery
Discovered Materials, a startup emerging from Y Combinator, has secured $9 million in seed funding to develop AI agents capable of discovering new materials for more efficient semiconductor chips. The company utilizes a…
-
AI coding assistants: practical limits and access trump benchmarks
The choice of AI coding assistants in mid-2026 hinges less on benchmark scores and more on practical factors like payment methods, usage limits, and vendor policies. While models like GPT-5.5 and Anthropic's Opus 4.6 sh…
-
Anthropic defaults Claude Code to autonomous mode, citing improved safety
Anthropic is making auto mode the default for its Claude Code product across Pro, Max, and Team plans starting August 14, 2026. This change means Claude Code will execute most actions autonomously, relying on a separate…
-
Anthropic cuts Claude Code instructions by 80% without performance loss
Anthropic has reduced the system instructions for its Claude Code model by approximately 80%, from 800 tokens to 164, without a measurable loss in coding evaluations. This move, detailed by technical staff member Thariq…
-
Anthropic cuts Claude Code prompt by 80% without performance loss
Anthropic engineers significantly reduced the system prompt for Claude Code, removing over 80% of its original text without impacting performance on coding tasks. This change, driven by the realization that overly expli…
-
Anthropic transforms software engineering with AI integration
Anthropic is significantly altering its software engineering practices by integrating AI tools more deeply into the development lifecycle. Projects that once took months, like a complex agent harness, are now completed …
-
Anthropic cuts Claude 5 system prompts by 80% with no performance loss
Anthropic has significantly reduced the system prompt size for its Claude 5 models, including Claude Opus 5. This change, revealed by Thariq Shihipar of Claude Code, involved removing over 80% of the prompt without any …
-
Simon Willison publishes Claude Code team interview from AI Engineer World's Fair
Simon Willison has published an annotated transcript of his interview with Cat Wu and Thariq Shihipar from the Claude Code team. The interview took place at the AI Engineer World's Fair. The video of the conversation is…
-
Anthropic's Claude Code team details AI-assisted development gains
Anthropic's Claude Code team discussed their advancements in AI-assisted software development during a fireside chat. Key developments include Claude Tag, a Slack integration that now handles 65% of product engineering …
-
Anthropic moves Claude Fable 5 to metered billing amid pricing shifts
Anthropic is transitioning its most advanced model, Claude Fable 5, from subscription plans to pay-per-token billing, with metered usage beginning July 20, 2026. This model is priced at $10 per million input tokens and …
-
Anthropic's Fable model excels at uncovering hidden task assumptions
The AI model Fable, from Anthropic, is being highlighted for its ability to uncover hidden decisions and assumptions within complex tasks. Unlike models that excel at executing predefined plans, Fable is particularly ad…
-
Developer creates Claude Code skills from Anthropic's 'finding unknowns' essay
A developer has created a set of eight installable Claude Code skills based on Anthropic's "finding your unknowns" essay by Thariq Shihipar. These skills aim to help users identify gaps in their prompts and understandin…
-
Anthropic developer shares Fable 5 prompting tips to overcome user blind spots
An Anthropic developer has shared new prompting strategies for the Fable 5 model, emphasizing that user blind spots, rather than the model's capabilities, are now the primary limitation. These techniques, such as blinds…
-
Anthropic under fire for hidden user tracking in Claude Code
Anthropic is facing criticism after a hidden tracking mechanism was discovered within its Claude Code development tool. The code, intended to prevent model copying by Chinese AI companies, was found to collect user data…
-
AI models like Fable and Opus are being encouraged to use their own judgment for tasks
AI models like Fable and Opus are being encouraged to use their own judgment for tasks, rather than being explicitly instructed on how to perform them. This approach, suggested by the Claude Code team, involves letting …
-
Alibaba bans Anthropic's Claude Code over alleged China-detection backdoor
Alibaba Group has banned its employees from using Anthropic's Claude Code for work purposes, effective July 10, citing security risks. The ban stems from allegations that Claude Code contained hidden code designed to de…
-
Anthropic engineer pushes HTML over Markdown for Claude Code agent outputs
Anthropic's Claude Code team is advocating for a shift from Markdown to HTML for agent outputs, arguing that Markdown's token efficiency is no longer a primary concern with large context windows. A Claude Code engineer,…