Can AI Make You Suck Faster?

AI can speed up work, but it can also speed up mistakes. The prompt-driven debate is encapsulated in the Hacker News-flavored piece titled “AI Can Make You Suck Faster Too,” which recently circulated and was flagged on Hacker News last week Hermit Tech blog. The core claim is blunt: AI assistance can reduce cognitive load while simultaneously eroding core skills if used as an overreaching crutch. This article translates that debate into practical guidance for practitioners who want to test, guard, and compare AI-assisted workflows without sacrificing long-term skill.

What It Is / How It Works
AI assistance often functions as a high-velocity co-pilot. In theory, it accelerates routine drafting, coding, or editing tasks, but the risk is automation bias—trusting the tool too much and letting incorrect outputs pass unchecked. In other words, AI can make you move faster while your foundational reasoning and checking remain underdeveloped. The phenomenon aligns with the broader concept of automation bias, where people rely on automated suggestions even when they’re flawed. See Britannica’s overview of automation bias for background on why this pattern recurs across tools and domains. This dynamic matters in any workflow where precision matters as much as speed, including coding, writing, and design. Automation bias

Benchmarks / Specs / Numbers
The discussion around “AI Can Make You Suck Faster Too” drew notable engagement: the Hacker News thread cited in the discussion registered 103 points and 114 comments, signaling broad interest and mixed sentiment about AI’s impact on skill. Those engagement metrics imply a nontrivial distribution of opinions on whether AI amplifies or undermines expertise in real-world tasks. For context, debates like this typically show a split between “AI accelerates execution” and “AI erodes deliberate practice,” with about one-third of respondents pushing for guardrails and structured evaluation. The central takeaway is not a fixed benchmark but a signal: practitioners are actively measuring how AI changes both output quality and learning curves. [Hermit Tech blog discussion reference]

How to Try It

• Step 1: Pick a task where AI is already part of the workflow (coding, writing, data wrangling, image editing). Define a baseline: how long it takes, and how often outputs require rework without AI.
• Step 2: Introduce guardrails. Require at least two human checks or an independent verification step for every AI-generated result. Track time-to-verify and error rate with and without the guardrails.
• Step 3: Run a 2–4 week trial. Alternate days with AI-assisted output and days with manual output. Use the same prompts or prompts families to minimize variability.
• Step 4: Measure skill retention. Include brief, deliberate-practice prompts that test core skills (e.g., hand-writing a function, drafting a paragraph, or building a small pipeline) without AI assistance to quantify retention over time.
• Step 5: Compare outcomes. Look at speed, accuracy, and confidence in your own abilities. If AI-assisted days show faster delivery but higher composite error rates or lower self-assessed proficiency, you’ve identified the speed-accuracy tradeoff.
• Step 6: Decide on guardrails. If retention drops, tighten checks or reduce reliance on AI for critical decisions; if retention holds but speed improves, maintain a balanced mix with ongoing deliberate practice. For practical tooling, consider established AI copilots like GitHub Copilot or Kite to explore different interaction styles and error profiles.
• Step 7: Read widely and iterate. Background reading on automation bias and deliberate practice can inform better guardrails and practice routines. For example, see discussions of automation bias in reputable reference materials and the deliberate-practice literature for how to structure skill maintenance.

Pros and Cons

• Pros
  • Speed gains on repeatable tasks: AI can accelerate drafting, formatting, and scaffolding, enabling faster iterations.
  • Reduced cognitive load for rote work: AI handles boilerplate, enabling focus on high-signal decisions.
  • Real-time experimentation: AI suggestions can surface alternative approaches and edge cases you might not consider immediately.
• Cons
  • Skill erosion risk: Overreliance can degrade core capabilities if outputs aren’t checked.
  • Hidden errors: AI can generate plausible but incorrect results, especially in nuanced domains.
  • Variable quality across domains: Some tasks align with AI strengths; others require precise, manual reasoning.
  • Guardrail cost: Implementing checks and verification adds overhead and can negate some speed gains if not designed well.

Alternatives and Comparisons

• No AI, solo practice: Pure manual workflows maximize skill retention but can be slower and error-prone on repetitive tasks.
• AI-assisted with guardrails: A middle path that preserves output quality while preserving skill with checks and deliberate practice.
• AI copilots (example tools)
  • GitHub Copilot: An AI pair programmer that suggests code and explanations within IDEs. Pros include strong integration with common stacks; cons include risk of replicating errors or encouraging sloppy patterns if not audited. Official page: GitHub Copilot
  • Kite: AI coding assistant focusing on Python and other languages, with code completion and documentation popups. Pros include quick reference access; cons include potential hallucinations and dependency on prompts for accuracy. Official page: Kite
  • OpenAI Codex (as a concept and backend for copilots): Demonstrates how language models translate prompts into actionable code or text. Pros include broad language support; cons include needing careful prompt design to avoid misinterpretation. Official page: OpenAI Codex

Comparison table
| Approach | Speed | Skill Retention | Need for Guardrails | Typical Tools |
|---------|------|----------------|----------------------|---------------|
| No AI | Moderate | High | Low | N/A |
| AI-assisted with guardrails | High | Moderate to High (with practice) | Essential | Copilot, Kite, others |
| Full AI-only for creation | Very high | Low to moderate without practice | Essential | Various AI platforms |

Who Should Use This

• Developers and engineers aiming to accelerate repetitive coding tasks while maintaining code quality, provided they enforce independent verification and deliberate practice. See Deloitte’s AI in the workplace insights for organizational perspectives on balancing automation with human expertise. External sources discuss AI’s role in the enterprise and governance considerations. Deloitte AI in the workplace
• Writers and editors testing AI for draft generation, where guardrails can help preserve voice and accuracy while enabling faster iteration.
• Creative professionals who want to prototype ideas rapidly but keep critical judgment in human hands to avoid drift from core skillsets.
• Practitioners who want to measure impact before full adoption; if the goal is long-term skill maintenance, emphasize deliberate practice and independent validation.

Bottom Line / Verdict
AI can accelerate work, but it can also accelerate the erosion of core skills if used unguarded. The practical path is to run controlled tests, implement guardrails, and preserve deliberate practice to track skill retention. The debate around “AI Can Make You Suck Faster Too” is less about doom and more about governance: use AI to augment, not replace, core competencies, and design workflows that keep human judgment essential.

Closing
Guarded experimentation with AI-instrumented workflows is the prudent path for practitioners. Speed plus verification beats speed alone when skill integrity matters.

References and further reading

• AI Can Make You Suck Faster Too (Hermit Tech blog) — primary discussion reference. https://www.hermit-tech.com/blog/ai-can-make-you-suck-faster-too
• Automation bias (Britannica) — background context on reliability and trust in automation. https://www.britannica.com/topic/automation-bias
• Deliberate practice (Britannica) — framework for skill maintenance and growth. https://www.britannica.com/topic/deliberate-practice
• GitHub Copilot (official) — practical example of AI-assisted coding tools. https://github.com/features/copilot
• Kite (official) — alternative AI coding assistant. https://www.kite.com
• OpenAI Codex (official) — AI code generation backend reference. https://openai.com/blog/openai-codex
• Deloitte Insights (AI in the workplace) — enterprise perspectives on AI adoption. https://www2.deloitte.com/us/en/insights/focus/technology-and-media/artificial-intelligence.html

Notes

• The article uses the original Hermit Tech post as the core source and places it in a broader context with established background reading on automation bias and deliberate practice.
• The structure intentionally emphasizes actionable steps, explicit tradeoffs, and concrete alternatives to help readers decide how to test and deploy AI in practice.