Key Takeaways
- Alex Lieberman lays out a vision to "kill and rebuild" consulting with a full-stack partnership model and an agent-first SDLC called MetaHarness.
- Enterprise AI adoption splits into single-player and multi-player modes; the latter offers exponential value but requires painful process reengineering.
- Arvid Kahl notes that MCP and API support is now a baseline requirement for software tools, and AI agents demand users with deep systems and domain knowledge to be effective.
- levelsio criticizes Apple's decade-long developer neglect and zero AI investment, predicting a decline similar to Windows in 2013, and says he's seriously considering moving to Linux and Android.
- levelsio releases GoldenPad, an open-source project bringing N64 GoldenEye 007 to iPhone/iPad natively via static recompilation, with iPad split-screen multiplayer in development.
- acquire.com reports 2026 M&A attractiveness will depend less on growth and more on profitability, retention, clean financials, founder independence, and a clear AI narrative.
1. AI Consulting and the New Workflow Stack
Alex Lieberman unveiled his strategy for "killing and rebuilding" the consulting industry: 2026 consultancies must become full-stack partners that embed strategy, change management, and internal tooling, rather than delivering 150-page decks. For many Fortune 500 firms, the first step is teaching executives to use Claude CoWork/Claude Code or Codex, with the goal of reducing friction to near zero. — via 1
He frames the goal as a "modern Bell Labs": instead of raising hundreds of millions on day one, the consultancy would fund an applied AI lab by solving difficult AI problems for large companies, repositioning consulting as "growth and innovation as a service." This model aims to generate cash flow while learning real pain points and building A-level teams that deliver tangible ROI. — via 1
AI applications split into single-player and multi-player modes: the former (e.g., swapping Google for GPT/Claude) requires zero behavior change and can generate $5M–$10M token agreements, but no compounding; the latter reshapes cross-functional processes and produces exponential value, though it is chaotic and painful. Most companies should start with single-player, but they won't see compounding returns. — via 1
His MetaHarness SDLC replaces Agile's ritual-led coordination with human+agent collaboration where agents are primary on-the-ground workers. It emphasizes a written-first culture (Markdown briefs), allowing agents to handle larger blueprints than Agile ceremonies, and uses specific Markdown artifact types to keep agents synchronized with the right context while minimizing entropy and drift in long tasks. — via 1
2. Enterprise AI Tooling: Integration and Skills
Arvid Kahl reports that enterprises now expect every software tool to provide MCP and API interfaces—he sees these requests on every deal at Podscan. This signals a shift where customers want to drive your service from their existing tools, rather than adopt a vendor-provided agent. — via 1
Using AI agents without understanding how software systems work quickly hits a productivity ceiling after a few hours, Kahl notes. Configuration, clearly articulating build requirements, and setting up validation/testing demand both programming knowledge and real domain expertise. — via 1
Ben Tossell criticizes Claude for not supporting local session creation from mobile, arguing this flaw pushes users toward Codex. — via 1
3. Apple's Developer Neglect and Open-Source Response
levelsio argues Apple is approaching a turning point similar to Windows in 2013: more than a decade of ignoring developers, buggy iOS/macOS releases (e.g., black camera freezes, broken Spotlight search), a lack of attention to detail, and zero AI investment are eroding its moat. He says he is seriously considering switching to Linux and Android for the first time. — via 1 2
He released GoldenPad, an open-source project that runs N64 GoldenEye 007 natively on iPhone/iPad—not via emulation but through static recompilation. He is also developing local split-screen multiplayer for iPad and plans macOS support later. — via 1
4. M&A Market Signals for 2026
acquire.com reports that going into 2026, company attractiveness in M&A will depend less on raw growth and more on profitability, retention, clean financials, operations independent of the founder, and a clear AI narrative. — via 1
New listings on acquire.com show active SaaS/e-commerce deal flow: a white-label SaaS for agencies with 5,000+ customers and ~$676K–$717K TTM revenue is priced at $336K; an enterprise employee advocacy SaaS with ~$1.08M TTM revenue (98% recurring) and 100 enterprise customers is priced at $2.6M; an organic castor oil e-commerce brand with $2.5M TTM revenue is asking $750K. These indicate a market where small profitable assets are trading at modest multiples. — via 1 2 3
