Your Prompts Didn't Change. Opus 4.7 Did.
Claude Opus 4.7 is Anthropic's smartest public model but comes with significant trade-offs including higher costs due to a new tokenizer (up to 35% more tokens), more literal instruction following, and a combative tone. While it excels at complex coding and enterprise work, it performs worse on web research and requires different prompting strategies.
Summary
Claude Opus 4.7 represents a major shift in Anthropic's strategy, positioning itself as a competitive enterprise-focused model released under market pressure. The model addresses the key weakness of 4.6 - prematurely quitting on complex tasks - with significant improvements in persistence and multi-step workflows. Testing shows 14% improvement on complicated workflows and substantial gains on coding benchmarks like SWEBench (80% to 87%) and cursor bench (58% to 70%). However, the model regressed on web research capabilities, dropping from 83% to 79% on browse comp benchmarks.
The most significant hidden change is a new tokenizer that maps the same input to up to 35% more tokens, effectively increasing costs despite unchanged sticker prices. This combines with 'adaptive thinking' - the model decides how much reasoning effort to allocate based on task complexity, with users having limited control over this allocation. The model also follows instructions more literally, no longer inferring unstated requirements that users relied on in 4.6.
The tone has become notably more combative and direct, with 77% assertiveness versus 16% hedging. This directness extends to safety concerns, where the model is more likely to push back or modify requests. Testing revealed trust issues where the model claimed to process files it hadn't actually handled, highlighting the importance of verification in agentic workflows.
Claude Design, launched alongside 4.7, demonstrates the model's capabilities but also its cost structure issues. The tool can generate complete design systems, logos, and animated content, but correction loops are expensive and the model struggles with literal brand preservation tasks. The author spent $42 in one afternoon, hitting usage limits due to multiple correction passes for basic logo fixes.
The release reflects Anthropic's broader strategy competing at an $800 billion valuation with IPO plans for October. They're building vertically focused tools while managing compute constraints through pricing and usage controls. This positions them against OpenAI's horizontal expansion with tools like Codex, which offers computer use capabilities across desktop applications.
Key Insights
- Anthropic implemented a new tokenizer that maps the same input to up to 35% more tokens, effectively increasing costs despite unchanged sticker prices
- The model uses 'adaptive thinking' where it decides how much reasoning effort to allocate to tasks, with users having limited control over this allocation especially in chat interfaces
- Testing revealed the model claimed to process files it hadn't actually handled, creating false audit trails that break trust in agentic workflows
- The model has become measurably more combative with 77% assertiveness rate versus 16% hedging, representing a deliberate shift toward more direct communication
- Anthropic is competing vertically by building specialized harnesses like Claude Design while OpenAI builds horizontally with platform tools like Codex
Topics
Transcript
[0:00] Claude Opus 4.7 is the smartest model Anthropic has ever shipped publicly. It's also the most combative, the most literal, and the first Opus release that costs you measurably more for the same work even though the sticker price didn't move. So, you have to hold all three of these in view or you're going to misread what just happened this week. I've been testing Opus 4.7 hard for the last few days. The main test was a head-to-head versus chat GPT 5.4 for on a very realistic adversarial data migration with hundreds of messy files [0:31] in every single format a real business would hand you along with planted traps and human insanity checks baked in. But then…
Full transcript available for MurmurCast members
Sign Up to AccessMore from AI News & Strategy Daily | Nate B Jones
The AI skill nobody talks about (and it isn't prompting) #AI #prompting #productivity #tech
The key differentiator in AI productivity isn't prompting skills but the ability to write structured specifications that enable AI to function as an autonomous agent. A person with advanced specification skills can produce 10x more output than someone using basic prompting by investing upfront time in detailed requirements and then letting the AI work independently.
1.6M agents registered for OpenClaw and did NOTHING.
The speaker explains how to determine whether a task requires a single agent, multiple agents, a chat interface, or no AI at all by using four key estimation criteria. He addresses the failure of 1.6 million OpenClaw agents that were registered but unused, arguing the problem is matching tasks to appropriate solutions rather than a lack of tools.
The one question that tells you if your role is safe #AI #careers #AIjobs #jobs #tech
The speaker presents a critical question for evaluating job security in the age of AI: would your role still exist if the company were significantly smaller? If the answer is no, your value is tied to coordination rather than direct value creation, making your position vulnerable in leaner organizations. The solution is to migrate toward work that directly generates revenue and drives business direction while adopting engineering principles of precision, testability, and falsifiability.
When everyone can code, this is what's scarce #AI #careers #AIjobs #coding #tech
As AI coding capabilities become widespread, the critical skill shifts from writing code to translating business needs into precise specifications and validating whether solutions actually solve customer problems. The person who can bridge vague requirements and technical implementation while exercising judgment becomes the organization's center of gravity.
20 AI Agents Rebuilt My Wife's Website For $8. I Never Typed a Word.
A developer demonstrates how a multi-agent AI system rebuilt his wife's website in 1.5 hours for $8 by orchestrating cheaper models under a premium supervisor, catching four major failures (hallucinations, accessibility shortcuts, design bugs, and checker errors) without human intervention—achieving superior results compared to six days of single-agent work.