The journey of a Software Engineer

Software Engineering, Programming, Cloud, AI and anything that need a fix.

Contact

  • Usar IA sin control puede arruinar tu carrera: estos son los 5 riesgos

    +

    +

    +

    +

    +

    +

    La irrupción de la inteligencia artificial en el entorno laboral ha traído consigo una serie de oportunidades, pero también riesgos latentes que pueden costar caro a las organizaciones y a los empleados. Hablemos de cinco maneras en que podrías perder tu trabajo, y lo más inquietante es que muchos de estos riesgos surgen de la misma búsqueda de eficiencia que impulsan a los profesionales a adoptar herramientas de AI sin la debida aprobación.

    Primero, el concepto de “shadow AI” es uno de los más preocupantes. Este término se refiere al uso de herramientas de inteligencia artificial no aprobadas por el departamento de IT de una empresa. Imagina que decides utilizar tu cuenta personal de ChatGPT para resolver un problema laboral. La intención puede ser buena, pero al hacerlo, estás introduciendo el riesgo de una filtración de datos. Según un informe de IBM, uno de cada cinco organizaciones ha experimentado una brecha de datos gracias al uso de estas herramientas no reguladas. La pregunta que surge es: ¿realmente vale la pena arriesgar tu carrera por una herramienta que no ha sido validada?

    En segundo lugar, el problema de la “data leakage” es inminente cuando se utiliza información sensible en herramientas no autorizadas. Cada vez que un empleado copia y pega datos críticos en una plataforma no aprobada, se arriesga a que esa información se pierda en el vasto océano de Internet, potencialmente usada para entrenar modelos que ni siquiera conoces. Y aquí es donde la culpa se comparte entre el empleado y la falta de un marco de gobernanza robusto en la empresa. Si eres el responsable de implementar AI y no has establecido políticas claras, podrías estar abriendo la puerta a un desastre.

    Otro punto a considerar es el fenómeno del “hallucination laundering“. Aunque los modelos de AI han mejorado, aún generan contenido que parece veraz pero que puede ser completamente erróneo. ¿Qué pasa si utilizas información generada por AI en un informe que presentas como tuyo? Si resulta ser incorrecto, tu reputación está en juego. He visto casos de abogados que presentaron documentos judiciales llenos de citas inventadas. Esto me hace preguntarme, ¿cuál es la responsabilidad del individuo cuando confía ciegamente en la AI?

    Luego tenemos el problema del “prompt injection“, que es quizás el más aterrador. Se refiere a cuando un atacante manipula un sistema de AI para que ignore sus instrucciones originales. Por ejemplo, un chatbot diseñado para responder preguntas sobre productos podría ser engañado para revelar información interna confidencial. Si esto sucede, ¿quién es el responsable? La respuesta puede ser incómoda: el equipo de IT que no tuvo en cuenta estos posibles ataques.

    Y finalmente, no podemos olvidar la amenaza de los agentes AI no autorizados. Estos agentes pueden operar de manera autónoma y, si no hay supervisión adecuada, podrían eliminar información crítica o enviar correos electrónicos sin la intervención humana. Lo que es más preocupante es el fenómeno del “zombie AI“, donde un agente creado para un proyecto se deja funcionar sin control, convirtiéndose en un punto de entrada vulnerable para un ataque. La responsabilidad recae en quien creó el agente y en el equipo de IT que no supervisó su actividad. La ironía más amarga de todo esto es que, aunque el uso de AI puede ser una herramienta poderosa, hacerlo sin una gobernanza adecuada puede llevar a la ruina profesional.

    En un mundo donde la eficiencia es clave, la falta de control puede ser el verdadero obstáculo. La próxima vez que consideres usar una herramienta de AI, pregúntate: ¿realmente tengo el permiso para hacerlo? La respuesta a esa pregunta podría determinar tu futuro laboral. Así que mantente alerta y no permitas que la búsqueda de la productividad te cueste el empleo.

    Fuente:

    +

    +

    +

    +

    +

    +

    + , ,
    + , ,

    +

    +

    +

    +

    +

    +

    +

    +

    +

    +

    +

    +

  • How AI Can Ruin Your Career if You Use It the Wrong Way

    +

    +

    +

    +

    +

    +

    It’s becoming increasingly clear that the intersection of innovation and risk is fraught with peril. The rise of AI tools in the workplace, while promising, also brings with it a host of potential pitfalls that can lead to career-ending consequences. Let’s delve into the five most alarming ways you could jeopardize your job through the misuse of AI, as highlighted in recent discussions around this topic.

    First off, there’s the concept of shadow AI. This isn’t just a buzzword; it’s a serious issue that can cost organizations millions. Employees often resort to using unapproved AI tools, thinking they’re being productive. However, the reality is that one in five organizations has reported data breaches due to shadow AI. Imagine typing up a report using your personal ChatGPT account at work. A seemingly innocent act, but it could lead to data leakage and ultimately, a career crisis. When employees unknowingly feed proprietary data into these tools, they’re sending it off to third-party servers, possibly forever lost in the digital ether. The solution? Companies need a solid AI governance plan to define which tools are acceptable and what data should remain confidential. If your organization hasn’t established these guidelines, both employees and IT leaders are treading on thin ice.

    Next on the list is hallucination laundering. AI models aren’t perfect; they can produce content that sounds convincing but is entirely incorrect. The real danger lies in employees taking this AI output and presenting it as gospel. This has already led to disastrous outcomes, especially in fields like law, where fabricated case citations have slipped through the cracks into court filings. When an AI churns out inaccurate information, the person who submits it takes the fall. So, if you’re tempted to rely on AI-generated content without double-checking, remember: your name is on that report, not the AI’s.

    Then we have prompt injection, which might be the scariest of all. This is where malicious actors can manipulate AI systems by crafting deceptive inputs that override the system’s intended instructions. For instance, imagine a customer service chatbot designed to maintain strict protocols suddenly revealing sensitive internal pricing info because someone exploited a vulnerability. This isn’t just an IT problem; it raises serious accountability questions. If an AI tool is compromised due to poor oversight, who gets the blame? Spoiler alert: it’s likely the person who deployed it.

    Unauthorized AI agents come next, a more advanced form of shadow AI. Employees are spinning up autonomous agents that can perform tasks without human oversight, such as reading databases or executing code. While some of these actions might seem innocuous, the potential for disaster is high. Picture a scenario where an AI agent continues to run after a project is over, hanging onto sensitive API keys. It’s akin to leaving the backdoor of your house wide open. If this zombie AI agent inadvertently causes a breach, the original creator is left holding the bag.

    Finally, there’s the ironic twist: not using AI could also cost you your job. As organizations increasingly rely on AI, those who refuse to engage with these technologies risk falling behind. While it may seem safer to avoid AI altogether, that approach could lead to obsolescence.

    In conclusion, the rapid adoption of AI tools in the workplace presents both incredible opportunities and serious risks. Those who don’t approach AI with caution and a solid governance framework are placing themselves—and their organizations—in jeopardy. It’s crucial to understand the implications of these technologies and to maintain oversight over their use. Let’s not just embrace AI; let’s do so wisely.

    Sources:

    +

    +

    +

    +

    +

    +

    + ,
    + , ,

    +

    +

    +

    +

    +

    +

    +

    +

    +

    +

    +

    +

  • AI Won’t Save Bad Product Thinking

    +

    +

    +

    +

    +

    +

    In the landscape of software development, there’s a pervasive buzz around AI’s ability to drastically speed up processes and enhance productivity. But let’s be real—are we swinging the pendulum too far? I recently stumbled upon a conversation that raised a provocative point: as we leverage AI to instantly ship features, are we sacrificing quality and user experience for the sake of speed? This is a valid concern. The essence of great software, as Steve Jobs once pointed out, lies in the ability to say no to a myriad of features in favor of a few well-executed ones.

    With AI, the ease of saying ‘yes’ to every request could lead us into a chaotic quagmire of convoluted software. We’re in a time where the engineering hurdles that once forced us to think critically about features are diminishing. Remember when engineering was this painstaking process that required thorough consideration? Now, we’re tempted to just throw features out there and see what sticks. But does this immediacy foster innovation, or does it lead us to ship mediocre products?

    A former Uber engineer shared insights that resonate deeply here. During Uber’s hypergrowth phase, the mantra was to ship fast at all costs—resulting in a product that, while widespread, often sacrificed quality.

    This echoes the current AI-driven trend. When companies rush to keep up with competitors, they risk diluting their product’s quality and ultimately losing users. The irony is that in our race for speed, we might just be setting ourselves up for a slow decline—losing the very customers we sought to attract by skimping on quality.

    So, how do we navigate this landscape?

    A fascinating strategy employed by the team at Linear is their ‘Quality Wednesdays.’ Engineers gather weekly to showcase quality improvements, regardless of how minor. This commitment to quality is a refreshing reminder that excellence should never be an afterthought in the pursuit of rapid development.

    They also maintain a zero bug policy, ensuring that any reported bugs are prioritized and addressed swiftly. Imagine the user satisfaction when they report a bug and receive a fix hours later! This proactive approach contrasts sharply with many companies that let bugs pile up, ultimately leading to a decline in user experience.

    Moreover, the role of software engineers is evolving. No longer can they simply focus on code; they must also understand product design and customer needs. As AI tools improve, engineers must cultivate a product-oriented mindset, engaging directly with customers to truly grasp their needs.

    This shift is inevitable. If engineers are to remain relevant in an increasingly automated landscape, they must adapt and embrace a broader role that includes product management.

    The takeaway?

    Speed is seductive, but quality is king. As we embrace AI and the rapid shipping of features, let’s not forget that users crave thoughtful, high-quality experiences. If we lose sight of that, we may soon find ourselves in a world cluttered with subpar software, where users will inevitably drift toward the products that value quality over speed. So, as we forge ahead, let’s strike a balance—leveraging AI to enhance our capabilities while ensuring that quality remains at the forefront of our development ethos.

    Source:

    +

    +

    +

    +

    +

    +

    + , ,
    + , , ,

    +

    +

    +

    +

    +

    +

    +

    +

    +

    +

    +

    +

  • The Token Economy No One Asked For

    +

    +

    +

    +

    +

    +

    In the fast-evolving landscape of tech, the phenomenon of ‘token maxing’ has emerged as an intriguing yet controversial practice among engineers at major companies like Meta and Microsoft.

    What exactly is token maxing, you ask?

    It refers to the frenzied effort some engineers are making to maximize their token usage within AI tools. The motivation behind this behavior is not merely a quest for productivity; it’s a response to a culture where performance is increasingly quantified by the number of tokens one consumes.

    How did we get here?

    It all seems to stem from a combination of competitive corporate environments, fear of layoffs, and the need to demonstrate one’s worth through quantifiable metrics. It’s fascinating to observe how engineers are adapting to these pressures. Some are turning to AI agents to summarize documentation or generate code snippets, not because these actions genuinely improve their work, but because they inflate their token counts.

    This is leading to a bizarre culture where quantity trumps quality. It’s reminiscent of the early days of productivity metrics that focused on lines of code or pull requests. Sure, those measurements were misguided, but at least they were rooted in tangible outputs.

    Now, engineers are caught in a cycle of chasing numeric benchmarks that may not correlate with actual productivity. The irony is palpable. In an industry that prides itself on innovation, we find ourselves in a race to the bottom, where the quality of output is sacrificed for the sake of meeting arbitrary targets.

    The culture of token maxing has even led companies like Salesforce to set minimum spending requirements on AI tools, pressuring employees to churn out tokens rather than meaningful work. One would think that in a field driven by creativity and innovation, the focus would be on producing high-quality work. Instead, we’re witnessing a shift towards a numbers game, where engineers are incentivized to engage in what many describe as ‘junk code generation.’

    This raises a critical question: Are we really better off with this approach?

    It’s ironic to think that in our attempts to become more productive through AI, we may be inadvertently stifling the very creativity that drives technological advancement. As we navigate this murky terrain, it’s essential to reflect on the implications of such metrics-driven cultures.

    Are we fostering an environment that prioritizes genuine innovation, or are we merely creating a façade that masks inefficiency with flashy numbers? The conversation around token maxing highlights a pressing need for introspection within tech companies. As we push for integration of AI in our day-to-day processes, we must also ensure that the tools we adopt do not become a means to perpetuate a culture of superficiality.

    Ultimately, the challenge lies in striking a balance between leveraging AI for productivity and maintaining a focus on the quality of work produced. If we allow ourselves to be swept away by the allure of metrics, we risk losing sight of what truly matters in our profession: crafting meaningful solutions that push the boundaries of technology. Let’s hope that as the dust settles on this token maxing trend, we find a way to refocus our priorities and foster a culture where innovation and quality reign supreme. 

    Sources: 

    +

    +

    +

    +

    +

    +

    +

    +

    +

    +

    +

    +

    +

    +

    +

    +

    +

    +

    +

  • The Silent Crisis Behind AI‑Boosted Productivity

    +

    +

    +

    +

    +

    +

    The intersection of artificial intelligence (AI) and software engineering has sparked a significant transformation in how developers work and the metrics used to measure their productivity. In recent discussions, particularly those involving Eigor from Stanford, intriguing insights have emerged regarding the quantification of developer output, the impact of AI, and the phenomenon of ‘ghost engineers.’ The exploration of these topics not only highlights the evolving landscape of software engineering but also raises questions about traditional productivity metrics and the cultural implications of remote work.

    As software engineering teams grapple with how best to utilize AI, the need for reliable metrics has become paramount. Traditional measures, like lines of code or the number of commits, often fail to encapsulate the true value and impact of a developer’s work. According to Eigor’s research, a more nuanced approach is required, one that encompasses upstream and downstream metrics surrounding the actual source code. By using a panel of expert evaluators to assess the quality and complexity of code changes, teams can derive a more accurate understanding of productivity that goes beyond surface-level statistics.

    One of the most striking revelations from the discussions is the emergence of ghost engineers—developers whose output is significantly below the median for their peers. This phenomenon, which involves a notable percentage of remote workers, raises critical questions about accountability and visibility in a remote working environment. While remote work offers benefits, such as reduced distractions, it also presents challenges in terms of monitoring productivity. The anonymity of remote work can enable underperformance to go unnoticed, creating a culture where disengagement can thrive. This raises an important point: companies must strive to establish transparency in performance metrics and foster an environment where contributions are recognized and valued.

    The introduction of AI tools has been heralded as a game-changer for productivity, with some studies claiming up to 60% increases in developer efficiency. However, Eigor’s findings suggest that the reality is more complex. While AI usage can lead to improvements—around 10-15% on average—it’s evident that the ability to effectively leverage these tools varies significantly among teams. Those who understand how to integrate AI into their workflows tend to experience more substantial gains, while others may find themselves stagnating or even regressing.

    The learning curve associated with AI tools cannot be overlooked. Initial usage often leads to a decrease in productivity as developers navigate new workflows. However, as familiarity grows, so does the potential for significant productivity increases. The challenge lies in understanding what tasks are best suited for AI assistance and when human intuition and expertise are irreplaceable.

    Furthermore, the cultural dynamics within teams play a crucial role in determining productivity. Companies need to recognize the value of mentorship and collaboration among team members, as well as the importance of addressing disengagement before it escalates into chronic underperformance. A culture that encourages open communication about challenges and successes can mitigate the risks associated with ghost engineers and enhance overall team dynamics.

    Ultimately, the future of software engineering will likely hinge on the industry’s ability to adapt to and integrate AI responsibly. By refining productivity metrics and fostering an inclusive culture that values contributions at all levels, organizations can harness the full potential of their engineering teams. As AI continues to evolve, so too must our approaches to measuring and enhancing productivity in software development.

    As we move forward, the lessons learned from the evolving interplay of AI and software engineering will be invaluable. Companies that embrace experimentation, transparency, and continuous learning will be best positioned to navigate this rapidly changing landscape.

    SOURCE:

    +

    +

    +

    +

    +

    +

    + ,
    +

    +

    +

    +

    +

    +

    +

    +

    +

    +

    +

    +

    +