Posts

Why 90% Accurate Still Feels Broken—And Why AI's Next 5% Changes Everything

Image
Deep Dive Dialogue The March of Nines: Why Cutting AI Error Rates Feels Like Compound Interest Hosts Bob & Elena unpack Google's shattered release cadence, the next six months of multi-file coding, Lusser's Law, and the Speech Recognition cliff. Part 1 — How the Six-Week Floor Shattered Bob Welcome back. Today we are looking at a rule everyone in AI took as gospel just a few months ago—that frontier model releases could never come faster than once every six to eight weeks. That ceiling just got blown to pieces. Elena Completely shattered. Look at Google's calendar over the last five weeks: Gemini 4 Argon dropped October 13, Carbon on November 10, Hydrogen on November 12, and Neon on November 18. Four major family checkpoints inside a 36-day window. Bob Wait—how did that happen? The whole logic behind the six-week floor was physics and logistics: you have to do a fresh, end-to-end pret...

Why You Can't See Beyond The Singularity

Image
The Singularity is defined as when the rate of progress moves so fast that no one can see beyond it. It is the event horizon of human technological progress when it becomes unified, when the deduplication of solutions reaches the maximum limit it can reach. At that point, it becomes a single unified intent, the sum collective of all human and digital thought where whatever it "wills," for lack of a better word, becomes reality. Therefore, you cannot see beyond it, because reality itself changes and bends wherever it looks. The reason the doubling of knowledge rate has gone from linear to exponential is because with each technological breakthrough that reduced the amount of effort and time lost to the duplication of effort, the rate of progress increased. Once the ability to share solutions became fast enough so that the people working on them did not have to rebuild the skills to work on them from scratch (due to death), and the speed of propagation reached the point that ...

Will LLM's Trigger a Catastrophic Loss in Critical Thinking Skills?

Image
As Large Language Models (LLMs) rapidly become the default interface for coding, writing, and problem-solving, a troubling trend is emerging in the academic data: a surge of recent studies reveals a sharp, measurable decline in critical thinking, independent analytical reasoning, and long-term retention among students and professionals who routinely offload their cognitive heavy lifting to AI. While most discussions treat this phenomenon as simple intellectual laziness, the reality unfolding inside the skull is far more destructive. When you hand over the friction of complex analysis to a black box, you aren't just putting your problem-solving skills on standby—you are triggering an automated biological demolition crew that physically strips those capabilities out of your brain. When you stop practicing a complex thinking skill—such as speaking a second language, playing an instrument, or solving advanced mathematics—your brain actively dismantles the neural pathways supporting tha...

Robot Assisted Code Review: Why Deep Reasoning Just Killed the Manual Code Walk

Image
In a very weird but real way, modern extended-reasoning LLMs do for software engineering overall what Rust did for memory and resource management. Rust didn’t invent the discipline required to prevent double-frees or data races; it simply codified the rigorous mental accounting that C and C++ systems programmers had to maintain in their heads, turning a fragile mental model into deterministic compiler enforcement. What an LLM running on extended reasoning mode does is act as a semantic borrow checker for distributed application logic. The Semantic Borrow Checker By executing exhaustive control flow checks, tracing object conformance through deep multi-layer call chains, and pressure-testing failure modes, an immense number of bugs, silent race conditions, and architectural flaws are surfaced well before code touches a runtime environment. The manual "code walk" or "desk check" has historically been the most effective, battle-tested practice in a vete...

Using LLMs for Code Review: A Practical Cheat Sheet

Image
LLM Code Review Cheat Sheet With a top model like Argon in high thinking mode, you can instruct it to perform specific logic checks that eliminate the main tedium of diff checking: Control Flow: Check logic control flow across relevant modules, focusing on the most recent changes made. Regressions: Check for any code that may have been broken by the recent changes. Compatibility: Check for any existing code that is incompatible with the recent changes and needs to be updated. Swallowed Exceptions: Check for exceptions or errors trapped in nested scopes that do not bubble up properly (e.g., a catch block that suppresses an error). Object Contracts: Check for objects passed across the target call hierarchy that are not forwarded properly, and flag code that needs updating to handle the latest object structure. Async Execution: Check for any unawaited async methods. Timing & Initialization: Check for timing errors—especially code that needs to wait or...

Stop Clinging to the Craft: Modern LLMs Can Review Code Better Than You Can!

Image
The era of trivial single-file autocomplete is over. With the release of flagship reasoning models like Gemini 4 Argon , AI has crossed a major capability boundary: the ability to build and verify global execution graphs better—and far more tirelessly—than humanly possible. I made a post on X/Twitter about programmers still thinking that LLMs are only good for generation, but can't review. I pointed out that was true until the flagship models like Argon incorporate internal call tree evaluation before the Chain of Thought step. I recently ran Argon on a 9-file complex source code file suite that comprised a very tricky object-passing and relay race baton-passing scheme to accomplish web page scraping. I asked it to do a deep examination of every inconsistent parameter use, every swallowed exception that would not bubble up, and every logic flow error where during the call tree a piece of code is not...

The Coming AI Theranos: How VC Panic Is Engineering the Biggest Fraud in Tech History

Image
When an entire asset class realizes its traditional underwriting criteria have burned to ash, capital allocators don’t suddenly become cautious. They become desperate for a savior. The conditions for a catastrophic, multi-billion-dollar tech debacle aren’t just present—they are virtually identical to the run-ups to Theranos and FTX, but with two to three times the deployable capital scrambling for a home. As traditional software moats evaporate under rapid foundation-model iteration, traditional tech investors are terrified. They need someone to tell them that the laws of economics and computing have been suspended by a breakthrough only they are visionary enough to back. Here is the exact anatomy of how that disaster is already assembling itself. 1. The Precondition: Technical Due Diligence Has Collapsed Theranos worked because non-technical board members and retail executives didn’t understand microfluidics. The...