The AI bubble & the career move no one's talking about
Here is this week's digest:
Ask HN: Why Opus4.6 was silently removed from Claude Code?
Users initially believed Claude Opus 4.6 was silently removed after 4.7's release, leading to concerns about cost and transparency. However, it's still accessible by typing /model claude-opus-4-6 in the chat or finding it under 'More Models' in the UI. Many find Opus 4.7 less cost-effective, with reduced usage caps, leading to subscriber dissatisfaction and a search for alternatives like a '$20 Codex subscription.' The discussion highlights a broader debate on AI provider transparency and profit-driven decisions versus user experience.
Ask HN: Would you take a job programming VMS?
For experienced programmers facing AI disruption, niche legacy systems like VMS offer a strategic career path. These roles provide a "safe harbor" to leverage decades of experience, where judgment and caution are highly valued over speed. Demand for these specific skills is high but the talent pool is small, often leading to job security, pay bumps, and training opportunities. This allows professionals to productively wait out the initial chaotic evolution of AI before re-engaging with stable new technologies, if necessary.
Ask HN: Is Elon Musk Overrated?
When assessing high-profile figures like Elon Musk, it's productive to separate their prowess as a capitalist from their perceived status as a visionary or leader. While undeniable in wealth creation and pioneering new industries such as space launch and electric vehicles, critics highlight potential financial complexities, the sustainability of ventures like Starlink, and the influence of a broad lack of skepticism. A valuable approach involves scrutinizing financial realities beyond perceived market dominance and differentiating between business acumen and broader societal contributions.
Ask HN: Am I getting old, or is working with AI juniors becoming a nightmare?
Developers are grappling with AI's impact on code quality and junior skills. While AI can boost productivity for experienced seniors who critically review its output, juniors often produce bloated, unreadable code, raising concerns about skill degradation and future maintenance. A key fear is AI training on its own low-quality output, leading to a downward spiral.
To mitigate these issues, effective AI integration requires:
- Careful Prompting and Critical Validation: Ensuring developers understand and scrutinize AI output.
- Self-Testing Feedback Loops: Implementing mechanisms for AI to validate its own work.
- Senior Guidance: Experienced developers are crucial for code review, guiding juniors to reduce cognitive debt, and curbing unnecessarily defensive code.
- Focus on Fundamentals: Maintaining an understanding of core logic and programming principles remains vital.
Ask HN: What Makes AI a Bubble?
The discussion delves into whether the current AI boom constitutes a bubble, exploring contrasting viewpoints.
Key arguments for a bubble include:
- High Costs vs. Revenue: Many AI companies, especially those relying on large language models (LLMs), operate with very thin margins or even at a loss due to exorbitant compute costs, subsidized by investor funding rather than sustainable profits.
- Unproven Productivity: Despite hype, many sectors haven't seen significant, transformative productivity gains from AI, leading to a mismatch between expectations and real-world application.
- Hype-Driven Investment: A lot of investment is driven by the 'new flashy object' phenomenon, similar to past bubbles (e.g., dot-com, railroads), where companies merely claim to 'use AI' without clear business models.
- Fragile Pricing: The profitability of AI products often hinges on underlying model providers (like Anthropic) continually cutting token prices.
Conversely, arguments against a bubble highlight:
- Real Revenue & Growth: Many AI companies exhibit high, accelerating revenues, indicating genuine market demand.
- Future Potential: Valuations reflect estimated future potential and continuous model improvements, not just current contributions.
- Market Nuances: P/E ratios are generally lower than during the dot-com era, and major investments come from cash-rich tech giants who understand the technology.
Productive Arguments: The debate emphasizes that even world-changing technologies can experience bubbles, and sustained value requires not just growth but also a clear path to profitability and tangible productivity improvements beyond speculation.
Add a comment: