Quick wins for a faster PC:
Clear out junk files and repair common Windows errorsFree Scan →Scan for outdated or missing drivers - takes under a minuteDriver Scan →Repair Windows errors before they cause bigger problemsFix Now →The “H.I. First” manifesto argues that people—not AI—should set software goals, supply domain context, weigh trade-offs, and remain accountable for the result. Its headline claim that 99% of AI workflows fail is rhetorical, not a verified failure rate: the article provides no study, sample, or measurement to support that number.
What “H.I. First” means
In Masum Billah’s DEV Community article, “H.I.” means human intelligence. The idea is not to avoid AI, but to keep engineering judgment in charge of AI-assisted work. Billah presents generic, unreviewed output as an “algorithmic average” and argues that engineers add value by applying domain rules, system constraints, optimization goals, and judgment.
![The Microsoft Copilot Bible: [3 in 1] A Beginner's Guide to Harness AI in Office 365 to Automate Workflows, Generate Content, Analyze Data, and 10x Your Productivity](https://m.media-amazon.com/images/I/41+5Qv4pd2L._SL160_.jpg)
As an Amazon Associate I earn from qualifying purchases.
That is the author’s framework and prescription, not a result established by an experiment. It is useful as a responsibility model: delegate execution where appropriate, but do not delegate the definition of success or ownership of consequences.
Is it true that 99% of AI workflows fail?
No substantiated failure rate is given in the article. It does not identify a dataset, study method, population, or time period behind the 99% figure. NIST’s relevant publications also do not establish that percentage. Read it as headline rhetoric, not as a measured industry statistic.
#1 Best Overall
The article raises a broader concern about whether AI use can weaken code quality or strategic depth, but posing that question does not demonstrate that such a decline has occurred. A reader should assess a workflow on its own outputs and risks rather than infer a universal failure rate from the headline.
Which responsibilities should stay with people?
Define the problem and constraints
People should state the objective, relevant domain context, system boundaries, and requirements before asking an AI system to produce work. Without those inputs, an answer can be plausible while missing the conditions that make it correct for a particular project.
Choose among strategic trade-offs
AI can help explore options, but people must decide which trade-offs fit the project—for example, how requirements and constraints should shape an implementation. Billah’s point is to use human judgment for synthesis and direction rather than treating generated output as a decision-maker.
Verify and govern the result
Generated work needs review proportionate to its impact. A human should decide whether it meets requirements, whether its risks are acceptable, and whether it should be approved or released. That keeps accountability with the people and organization responsible for the system.
Rank #3
What work can be delegated to AI?
Billah’s proposed division is AI assistance with bounded implementation tasks and human control over context, strategy, and governance. In practice, a task is a better candidate for delegation when its scope is clear, its output can be checked, and errors can be caught before they create consequential effects.
- Give the AI a defined task, relevant constraints, and an expected output.
- Review the result against requirements rather than relying on plausibility or fluency.
- Test generated software and examine security implications before accepting changes.
- Keep a record of generated artifacts and who reviewed and approved them when traceability matters.
- Do not delegate a decision whose consequences the responsible team cannot evaluate or own.
These are practical applications of the manifesto’s human-led approach, not a guarantee that review will catch every defect. The amount of oversight should reflect the task’s sensitivity and consequences.
How NIST grounds the security side of the argument
NIST’s SP 800-218A, published in July 2024, adds practices for generative AI and dual-use foundation models to NIST’s Secure Software Development Framework. It provides a basis for considering AI-related security and secure-development practices; it does not validate the manifesto’s 99% headline or prove a particular workflow is safe.
Do these 3 things before closing this tab:
1Repair Windows errors before they cause bigger problems2Fix the driver behind crashes, sound loss and screen glitches3Clear out junk files and repair common Windows errorsNIST notes that “AI model and system development is still much more of an art than an exact science, requiring developers to interact with model code, training data, and other parameters over multiple iterations.” That observation supports treating AI-assisted development as work that needs iterative evaluation, not as a process in which generated output is automatically reliable.
The NIST National Cybersecurity Center of Excellence’s DevSecOps reference model offers a concrete example of human-directed AI assistance: it describes oversight, review gates, traceability, and accountable approval of generated artifacts. It is an institutional model, not evidence that human oversight prevents every failure or improves outcomes by a known amount.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.A practical test for an H.I.-first workflow
Before adopting an AI-assisted process, ask who directs the work and how its output will be checked:
- Objective: Has a person defined what success means?
- Context: Are the domain rules, system constraints, and relevant requirements available to the people directing the work?
- Delegation: Is the AI task bounded enough for someone to evaluate its output?
- Verification: Are the generated artifacts reviewed and tested before acceptance?
- Security and traceability: Are risks considered and artifacts traceable where the project requires it?
- Accountability: Is it clear who approves the work and owns the outcome?
If a workflow cannot answer who verifies the output or who accepts responsibility for it, adding more prompt detail does not resolve the underlying governance problem. The manifesto’s practical emphasis is to improve problem definition and architectural critique, not merely to keep refining prompts.
Crashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteWindows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallQuick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
