Opinionllmcoding agentssandboxingai security
Coding Agents Reshape Software Engineering And Tooling
8.0
On Jan. 8, 2026, a practitioner joined the Oxide and Friends podcast to offer one-, three-, and six-year predictions about LLMs and coding agents. He predicts that within a year LLMs (citing Claude Opus 4.5 and GPT-5.2) will produce clearly high-quality code, sandboxing will be solved, and a headline security "Challenger" event is likely; three-year forecasts foresee major productivity and tooling shifts.
Key Points
- 1Predicts LLMs will produce high-quality code within one year, citing Claude Opus 4.5 and GPT-5.2
- 2Highlights solved sandboxing and rising attack risk; warns current practices normalize deviance and invite disaster
- 3Suggests major productivity shifts—browser projects and coding roles will change, requiring secure, UX-friendly sandboxes
Scoring Rationale
Timely practitioner observations tied to recent model releases; limited by single-source predictions and anecdotal evidence.
Sources
Public references used for this report.
Practice with real Logistics & Shipping data
90 SQL & Python problems · 15 industry datasets
Used by DS/ML engineers at top companies
High-Value Overnight OrdersEasyDelivered International ShipmentsMediumOn-Time Delivery Rate by CarrierHard
250 free problems · No credit card
See all Logistics & Shipping problems