Predictions

Evidence so far

• Okta’s 2024 Businesses at Work report shows large enterprises averaging 200+ SaaS apps, and follow-up customer briefings now include formal ‘rationalization’ workstreams.
• ServiceNow, Salesforce, and Atlassian all highlighted SaaS consolidation programs on their 2024 earnings calls, framing vendor trims as an explicit margin lever for customers.
• Microsoft, Salesforce, and ServiceNow are trading bigger bundles (E5 + Security Copilot, Unlimited Suite, Now Assist Pro Plus) for multi-year commitments that remove dozens of point solutions—exactly the consolidation wave this prediction described.

Evidence so far

• PostgreSQL reached #1 most-used database globally (~45% of developers, Stack Overflow 2025), driven heavily by multi-model capabilities — pgvector, JSONB, PostGIS, TimescaleDB.
• Oracle 23c launched as a 'converged' multi-model database (row, columnar, JSON, spatial, vector). AWS Neptune, Azure Cosmos DB, and MongoDB Atlas all now market primarily as multi-model platforms.
• Elastic, Couchbase, and MongoDB shipped graph, vector, or columnar engines inside their primary products in 2024, validating the convergence thesis.
• Startups increasingly default to Postgres + extensions over polyglot stacks, but no single 'winner' has emerged as the definitive multi-model standard — the pattern is converging across many engines rather than consolidating into one.

Evidence so far

• Adoption arrived on time: 84% of developers used or planned to use AI tools in Stack Overflow's 2025 survey, and 51% of professionals used them daily; Google's 2025 DORA report put adoption at 90%.
• Orchestration went mainstream: in Stack Overflow's 2026 survey, 66% of developers using AI worked in coding agents, and the share working as an organization of one rose from 4% to 10%.
• Against 'breadth beats depth': the same 2026 survey found AI used mostly for code in familiar areas (67%) and debugging (61%), and only 20% for deploying or operating production systems.

Evidence so far

• Rust is now in the AWS SDK, Windows kernel, Android platform, Chrome infrastructure, and Linux kernel modules — mainstream adoption is no longer debatable.
• Go 1.23+ telemetry hooks and WASI investments make it the default for platform teams migrating latency-sensitive services from Python and Node.
• CISA and NIST memory-safety directives are steering federal procurement toward Rust, Swift, Kotlin, and TypeScript for all new development.
• Mojo and similar AI-native languages launched, reflecting demand for Python ergonomics at GPU speed — the AI ecosystem is spawning its own language category.

Evidence so far

• Bezos confirmed at the NYT DealBook conference in late 2024 that he's dedicating 95% of his time at Amazon to AI development, effectively returning to the frontlines as predicted.
• Amazon invested $8B in Anthropic and is building a supercomputer with its proprietary Trainium chips — the AI infrastructure bet is massive and accelerating.
• Amazon is developing 1,000+ internal AI applications and an ambitious multimodal model processing images, video, and text simultaneously.
• Bezos also launched Project Prometheus, a $6.2B AI startup focused on physical AI for aerospace and manufacturing — extending his AI ambitions beyond Amazon itself.

Evidence so far

• Go adoption continues to climb — its deliberate simplicity means a team of mixed experience levels produces remarkably uniform code. No inheritance, no generics abuse, one way to format.
• Rust's strict compiler enforces consistency at the language level. Two developers solving the same problem produce structurally similar solutions because the borrow checker leaves fewer valid paths.
• TypeScript's type system constrains JavaScript's anything-goes flexibility, bringing consistency to frontend codebases that previously varied wildly by developer preference.
• Java remains widely used but exemplifies the variance problem: the same feature can be a 50-line pattern-heavy implementation or a 15-line direct solution depending on who wrote it.