DEV Community: Gabriel Fernandes

Preparing Your Full-Stack Application for Success: Essential Scalability Strategies!

Gabriel Fernandes — Mon, 10 Mar 2025 15:26:40 +0000

Scaling Strategies for Full-Stack Applications

Introduction

As your full-stack application gains traction, its ability to handle increasing user traffic and data becomes paramount. Scaling ensures a smooth user experience and prevents performance bottlenecks. This article explores the two main scaling strategies: horizontal scaling and vertical scaling, their applications, benefits, and considerations for full-stack development.

Understanding Scaling Needs

Before implementing specific techniques, it's crucial to understand your application's scaling requirements:

Traffic Patterns: Analyze traffic surges during peak hours or specific events to determine if scaling should be dynamic or static.
Resource Utilization: Monitor CPU, memory, and storage metrics. Scaling becomes necessary when these resources consistently reach high utilization levels.
Application Architecture: Your architecture (monolithic vs. microservices) will influence your scaling approach. Microservices are generally easier to scale horizontally.
Cost Considerations: Factor in hardware upgrades, additional server instances, and potential licensing fees.
Performance Requirements: Consider desired response times and overall performance goals.

Horizontal Scaling: Distributing the Load

Horizontal scaling, or "scaling out," involves adding more servers to distribute workload across multiple machines. This approach is ideal when adding more processing power and handling increased traffic are primary concerns.

Benefits

Increased Capacity: More servers translate to increased processing power, memory, and storage.
Improved Fault Tolerance: If one server fails, others can handle the load, ensuring high availability.
Flexibility and Cost-Effectiveness: Add or remove servers based on actual demand.
Easier Management of Large Systems: Individual servers remain relatively simple to manage.

Implementation Techniques

Load Balancing: Distributes incoming traffic across multiple servers.
Database Sharding: Splits data across multiple servers based on pre-defined criteria.
Microservices Architecture: Each microservice can be independently scaled based on specific resource requirements.
Containerization: Technologies like Docker allow for easy deployment and scaling across multiple servers.

Considerations

Increased Complexity: Managing multiple servers adds complexity in monitoring, logging, and configuration.
Network Bandwidth: Can increase network traffic between servers.
Data Consistency: Maintaining consistency across servers requires careful planning.

Vertical Scaling: Powering Up a Single Server

Vertical scaling, or "scaling up," involves upgrading hardware resources of your existing server. This approach suits applications with predictable workloads or when adding more servers isn't feasible.

Benefits

Simpler Management: Maintaining a single server is generally easier than managing multiple servers.
Lower Initial Cost: Upgrading a single server can be less expensive initially.
Faster Implementation: Can be implemented more quickly than horizontal scaling.

Implementation Techniques

CPU Upgrades: Increasing cores or upgrading to a faster CPU improves processing power.
Memory Upgrades: Adding more RAM handles larger datasets and improves performance.
Storage Upgrades: Ensures sufficient space for growing data needs.
Database Optimization: Optimizing queries and indexes improves performance without hardware upgrades.

Considerations

Limited Scalability: There's a physical limit to how much you can upgrade a single server.
Single Point of Failure: Hardware failure results in complete downtime.
Increased Cost Over Time: Continuous hardware upgrades can become expensive long-term.
Resource Bottlenecks: Upgrading one resource might not address bottlenecks in another area.

Choosing the Right Strategy

The optimal scaling strategy depends on various factors:

Choose Horizontal Scaling When:

You anticipate significant growth in traffic or data volume
High availability and fault tolerance are crucial
Your application uses a microservices architecture
Cost-effectiveness is a primary concern

Choose Vertical Scaling When:

You have a predictable workload and user base
Simplicity and ease of management are priorities
Initial cost is a major constraint
Sufficient improvements can be achieved through hardware upgrades

The Hybrid Approach

In many real-world scenarios, combining both strategies is most effective. Start by vertically scaling to handle initial growth, then migrate to horizontal scaling when necessary to balance cost-effectiveness, performance, and manageability.

Best Practices for Implementation

Monitor and Analyze: Continuously track performance metrics to identify bottlenecks.
Automate Scaling: Implement solutions that dynamically scale resources based on predefined thresholds.
Test Thoroughly: Test scaling changes in a staging environment before production deployment.
Plan for Failure: Design with disaster recovery in mind, including backups and redundancy measures.
Security Considerations: Secure all communication channels between servers and implement robust access control.

Beyond Scaling: Additional Considerations

Caching: Reduce database load by caching frequently accessed data.
Content Delivery Networks (CDNs): Serve static content from geographically distributed edge servers.
Code Optimization: Regularly review and optimize application code.
Monitoring and Logging: Implement comprehensive solutions to identify issues and track resource utilization.

Conclusion

Understanding horizontal and vertical scaling allows you to develop a robust strategy that ensures optimal performance as your application grows. Scaling is an ongoing process—continuously monitor performance and adapt your strategy accordingly. Being proactive ensures your full-stack application scales seamlessly to meet user demands and delivers a consistently exceptional experience.

Get your Front End ready for Millions TODAY!🚀

Gabriel Fernandes — Mon, 10 Mar 2025 15:07:34 +0000

Introduction: The Growing Pains of a Frontend Codebase
Every frontend project starts small and simple—a few components, some state management, maybe a couple of API calls. Fast forward a year, and suddenly:

❌ Code reviews take forever
❌ Making a small UI change breaks unexpected parts of the app
❌ Deployments become stressful, debugging is a nightmare
❌ Onboarding new developers takes weeks

I’ve been through multiple frontend projects that hit these exact roadblocks. Looking back, there are things I wish I had done earlier to avoid the mess.

In this article, I’ll share 10 key lessons I’ve learned about scaling frontend codebases the hard way.

1️⃣ Keep the Codebase Modular From Day One
The mistake: Everything was dumped into a single components/ folder with zero structure.
The fix: Organize by feature, not type (e.g., features/dashboard/ instead of components/).
Takeaway: A modular codebase scales better, and refactoring later is painful.
✅ Best Practice:
📌 Use feature-based architecture
📌 Keep components small & reusable
📌 Separate UI, state, and logic early

2️⃣ Choose the Right State Management Approach Early
The mistake: We started with a custom state manager, which worked until it didn’t.
The fix: Use the simplest state management solution that fits your needs (Context API, Zustand, Redux, or React Query).
Takeaway: Overcomplicated state management is a tech debt time bomb.
✅ Best Practice:
📌 Keep local state inside components whenever possible
📌 Use React Query for server state instead of Redux
📌 Avoid prop drilling hell with Context or Zustand

3️⃣ CI/CD & Automated Deployments Are Not an Afterthought
The mistake: Manual deployments led to human errors and broken releases.
The fix: Automate the entire process (testing, builds, deployments) using CI/CD tools.
Takeaway: If you don’t automate early, you’ll regret it when your team grows.
✅ Best Practice:
📌 Set up NX or Turborepo for efficient builds
📌 Use feature flags to safely deploy unfinished features
📌 Run automated tests before merging

4️⃣ Performance Optimization Is Important
The mistake: We didn’t care about performance until users started complaining.
The fix: Optimize early by monitoring bundle sizes and network requests, note any unnecessary re-renders.
Takeaway: Performance bottlenecks are much harder to fix later.
✅ Best Practice:
📌 Use lazy loading & code splitting (React’s Suspense)
📌 Optimize bundle size with tree shaking
📌 Minimize re-renders with React.memo & useCallback

5️⃣ Proper Testing Saves You From Deployment Nightmares
The mistake: We relied on manual testing instead of automating tests.
The fix: A solid test strategy includes unit, integration, and E2E tests.
Takeaway: The best time to start writing tests was yesterday.
✅ Best Practice:
📌 Write unit tests for critical business logic
📌 Use React Testing Library for component tests
📌 Add Cypress or Playwright for E2E tests

6️⃣ Avoid Over-Engineering: Simple Is Better
The mistake: We overcomplicated everything with abstractions and unnecessary layers.
The fix: Build for today’s needs, but design with tomorrow in mind.
Takeaway: Complexity grows fast; keep it as simple as possible.
✅ Best Practice:
📌 Don't abstract too early—wait until patterns emerge
📌 Avoid unnecessary custom hooks if built-in hooks work fine
📌 Keep dependencies minimal to reduce tech debt

7️⃣ Documentation Is a Lifesaver
The mistake: The codebase relied on tribal knowledge, making onboarding painful.
The fix: Keep lightweight, useful documentation (not just API specs).
Takeaway: Well-documented projects are easier to scale.
✅ Best Practice:
📌 Use Storybook for UI documentation
📌 Maintain a simple README with architecture decisions
📌 Write code comments where necessary (but don’t overdo it)

8️⃣ TypeScript Will Save You From a Lot of Debugging
The mistake: Started with plain JS, then struggled with undefined errors everywhere.
The fix: Adopted TypeScript to reduce runtime errors.
Takeaway: Types catch issues before they become bugs.
✅ Best Practice:
📌 Use TypeScript from day one
📌 Type function props and API responses properly
📌 Keep types clean (avoid any at all costs!)

9️⃣ API Communication Should Be Standardized
The mistake: Different parts of the app handled API requests inconsistently.
The fix: Created a unified API service with consistent response types and error handling.
Takeaway: Standardized API calls reduce bugs & improve maintainability.
✅ Best Practice:
📌 Use React Query or SWR for fetching
📌 Centralize API logic in a single module
📌 Use error boundaries to catch failures gracefully

🔟 Scaling a Frontend Isn’t Just About Code; It’s About Process
At the end of the day, a scalable frontend isn’t just good architecture—it’s also about workflow, automation, and team collaboration.

If I could go back in time, I’d tell myself:
✅ Plan for modularization early
✅ Keep things simple
✅ Invest in automation, testing, and CI/CD before they become pain points

💡 What’s your biggest lesson from scaling a frontend? Let’s discuss!

Art and the New Technologies

Gabriel Fernandes — Tue, 19 Mar 2024 15:21:23 +0000

Continuing on my journey to understand the role of technology in the history of art, perhaps to try to grasp a bit more the natural progression of AI in this new stage of technology, or to broaden our horizons when discussing technological applications beyond what we may be accustomed to.

In the previous article, we discussed how photography revolutionized art and how impactful it was, motivating an entire generation of painters to "initiate" a new movement, no longer seeking to represent figures perfectly, but rather marked by vague contours and quick brushstrokes, now known as Impressionism.

Below, we can see what was considered the world's first permanent photograph:

Joseph Nicéphore Niépce, "View from the Window at Le Gras" (1826-27)

Still following that very similar model of large cameras and photos taken by professional photographers, a period where it was even common for postmortem photography to be taken if individuals had never been portrayed in life.

But by the end of the 19th century, cameras became increasingly portable, democratizing their use significantly, with artists using them to represent their works, document, and sell; thus promoting their work.

Thomas Eakins, "Study of Human Motion"

It was interesting to note that people who wanted to be models and portrayed simply had their photos taken, and painters were in charge of painting them, without the person needing to pose for hours. It was also common during these times for jewelry and accessories to be loaned to these artists for a more faithful representation.

Franz Xaver Winterhalter, "Eugénia de Montijo" (1853)

In the 20th century, particularly in contemporary art, photography also serves as a lasting record of ephemeral arts that naturally expire. Sometimes it portrays an artist performing an action they will not repeat forever; sometimes it captures interventions in nature, graffiti on condemned buildings, and many others... with the work having its perishable character in general.

The work goes away, but the record remains. (?)

Anne Imhof, "Faust" (2017)

Richard Long, "Sahara Circle" (1988)

Now, photography as an art form took a long time to become popular, finally achieving this around the 1980s with the so-called fine art photography.

Cindy Sherman, "Untitled #96" (1981)

It's interesting to think that color photographs date back to the mid-1870s, but they were only widely adopted almost 100 years later. For a long time, color photography was considered amateurish by various photographers and photojournalists, where contrast and light were more important for "True Photography", and colors could detract from these main characteristics. Something similar happened shortly after with the emergence of digital cameras, where professionals often chose to stick with their analog cameras, not embracing the new technology.

One of the most well-known photographers who transitioned from analog to digital was the Brazilian Sebastião Salgado.

Sebastião Salgado, photograph from the series "Amazônia" (2013-2019)

It's worth noting that each stage in this chronology did not compel everyone to change their way of making art or invalidate what came before. Each means of expression, whether photo, painting, or writing, is strictly linked to what the artist seeks to convey, and each carries its clear criteria.

Technology often does not destroy or replace art, and it's undeniable that it can change the way art is made, often expanding its horizons.

Much is said about photography as a protagonist of technology in the art world, but we must always remember that it was not the only one.

Video art, for example, derived from photography, also carries its permanent record character of something ephemeral, transmitting and preserving for the future.

In this regard, we can mention well-known artists who dabbled with and embraced technologies, such as Dali and Picasso, who also had their involvement in films.

Scene from Un Chien Andalou, by Salvador Dali and Luis Buñuel (1928)

In the 1950s, TV became increasingly common, and by the 1960s, we had the rebirth of Hollywood. Undoubtedly, this greatly contributed to this artistic model being widely consumed.

-
And after cinema, what we have most recently are the so-called Video Installations.

Works with modern features, as seen below, where the artist doesn't show just one video, but a series of them, displayed simultaneously on televisions, organized within the states that make up the map of the United States of America, outlined by neon lights.

Nam June Paik, "Electronic Superhighway: Continental U.S., Alaska and Hawaii" (1995)

Below, we see another example of a technological installation boasting lights and projections... First, because it's fun, and second, because it further demonstrates the artist's mastery of technology and the possibilities it promotes.

Janet Echelman, "Dream Catcher" (2001)

Imagine yourself in an empty room with white walls, you can already express yourself there in some way; now imagine you were given a pencil... now imagine you were given a brush and paint... now imagine you were given a camera, a camcorder... imagine you were given a sound device... imagine you were given a projector... Do you see how the possibilities follow technological advancements?

One very interesting thing is also to see how sometimes artists cannot realize an art they envisioned alone, and they need support from others to achieve this end, something very present in Contemporary Art.

What we see below is an example of this, an installation located at the Inhotim Museum, MG, Brazil. Where the artist Janet Cardiff had the support of a sound engineer to build this acoustic room, where strategically positioned speakers reproduce the voice of each member of a choir, singing "Spem in Alium" by the English composer Thomas Tallis.

Each speaker containing the voice of only one person reproduces the idea of being in the presence of the entire choir, especially when we approach a certain corner of this room, an installation that I have visited and highly recommend.

Janet Cardiff, "Forty Part Motet: (A reworking of 'Spem in Alium' by Thomas Tallis 1573)" (2001)

And, finally, I couldn't fail to mention the impact of technology on the study of art, which you are probably using now. It can show us at any moment a work of art of our choice in details and colors. Not long ago, this was something exclusive to books that often contained artistic representations only in black and white, and people had to imagine the colors.

And these were some of the examples I based on to discuss a bit about how technology emerges and contributes to art in various ways. I could even say a little more, citing more recent examples, such as mapped projection works, using projections on sculptures to simulate their movement; virtual reality arts, like those by artist Marina Abramovic; interactive, immersive artworks, and much more.

For me, it's very interesting to know all this and to be present in this world of possibilities. Because it shows how technology is not just what we use and debate as programmers when we see a new innovative framework or a more optimized solution for our specific needs. But rather, how versatile it is and how capable it is of promoting evolutions at both individual and collective levels, and it does so very well.

A Arte e as Novas Tecnologias

Gabriel Fernandes — Wed, 13 Mar 2024 16:20:03 +0000

Continuando na minha jornada em entender como foi o principio da tecnologia na historia da arte, talvez para tentar entender uma pouco mais o movimento natural das IA’s na nessa nova etapa da tecnologia, ou para expandir um pouco os horizontes para quando falarmos das aplicações tecnologicas irmos além do que possamos estar acostumados.

No artigo anterior falamos em como a fotografia reviolucionou a arte e quão impactante foi a ponto de motivar toda uma geração de pintores a “iniciar’ um novo moviemento, que não mais buscava representar as figuras de maneira perfeita, mas sim marcada por contornos imprecisos e pinceladas rápidas, hoje conhecido como Impressionismo.

Abaixo podemos ver aquela que foi considerada a primeira fotografia permanente do mundo:

Joseph Nicéphore Niépce, "Ponto de vista de Gras"(1826-27)

Ainda seguindo aquele modelo bem parecido de grandes câmeras e fotos realizadas por fotógrafos profissionais, período onde era até mesmo comum o registro fotográfico de pessoas já falecidas, caso elas nunca tivessem sido retratadas na vida.

Mas já ao final do séc 19, as câmeras se tornaram cada vez mais portáteis, democratizando bastante seu uso, onde artistas as usavam para representar suas obras, registrar e vender; divulgando assim seu trabalho.

Thomas Eakins, "Estudo de movimento humano"

Era interessante notar que as pessoas que queriam ser modelos e retratadas apenas tiravam suas fotos, e os pintores se encarregavam de pintá-las, sem que a pessoa precisasse ficar por horas posando. Nesses momentos, também era comum que jóias e adereços fossem emprestados a esses artistas para uma representação mais fidedigna.

Franz Xaver Winterhalter, "Eugénia de Montijo (1853)"

Já no séc XX, sobretudo na arte contemporânea, a fotografia também já serve como registro duradouro de artes efêmeras, que se esgotam naturalmente. Ora representando um artista executando uma ação a qual ele não ficará repetindo eternamente; ora representando intervenções na natureza, ora grafites em prédios condenados à demolição e muitos outros... tendo a obra seu caráter perecível no geral.

A obra se vai, o registro permanece. (?)

Anne Imhof, "Faust" (2017)

Richard Long, "Círculo do Saara" (1988)

Agora, a fotografia como arte por si só levou muito tempo para se tornar popular, conseguindo tal feito finalmente por volta dos anos 80, com a chamada fine art photography.

Cindy Sherman, "Sem título n96" (1981)

É interessante pensar que as primeiras fotografias coloridas datam de meados de 1870, porém só são amplamente adotadas quase 100 anos depois. Por muito tempo, a fotografia a cores era considerada, por diversos fotógrafos e fotojornalistas, amadora, onde contraste e luz eram mais importantes para a "Verdadeira Fotografia", e as cores poderiam atrapalhar o foco dessas características principais. Algo similar aconteceu um pouco depois, com o surgimento das câmeras digitais, onde profissionais muitas vezes optavam por permanecer com suas câmeras analógicas, não aderindo então às novidades.

Um dos mais conhecidos fotógrafos, que migrou da analógica para o digital, foi o brasileiro Sebastião Salgado.

Sebastião Salgado, fotografia da série "Amazônia" (2013-2019)

É válido ressaltar também que cada etapa nessa cronologia não fez com que todos mudassem sua maneira de fazer arte ou invalidou a que veio antes. Cada meio de expressão, seja ele foto, pintura ou escrita, está estritamente ligado ao que o artista busca transmitir, e todos carregam seus nítidos critérios.

A tecnologia muitas vezes não destrói ou substitui a arte, e é inegável que ela consegue mudar a maneira como a arte é feita; quando, muitas vezes, não expande seus horizontes.

Muito se fala da fotografia como protagonista da tecnologia no meio da arte, mas devemos sempre lembrar que ela não foi a única.

A videoarte, por exemplo, derivada da fotografia, também carrega seu caráter de registro permanente de algo efêmero, transmitindo e guardando para o futuro.

Nesse quesito, podemos citar artistas conhecidos que transitaram e flertaram com as tecnologias, como Dali e Picasso, que também tiveram suas participações em filmes.

Cena de Um cão andaluz, de Salvador Dali e Luis Buñuel (1928)

Nos anos 50 a TV se torna cada vez mais comum e em 60 temos o renascimento de Hollywood. O que sem duvida contribuiu e muito para que esse modelo artístico fosse consumido amplamente.

-
E após o cinema, o que temos de mais recente ainda, são as chamadas Vídeo instalações

Obras com fatores modernos, como vemos abaixo , o artista não mostra um vídeo só, mas uma série deles, que são mostrados ao mesmo tempo em televisores, organizados dentro dos estados que compõe o mapa dos Estados Unidos da América, contornados por neon.

Nam June Paik, "Eletronic Superhighway: Continental U.S., Alaska and hawaii" (1995)

Vemos abaixo outro exemplo de instalação tecnológica que se gaba de luzes e de projeções… Primeiro porque é divertido e segundo porque mostra ainda mais o domínio da tecnologia pelo artista, e as possibilidades promovidas pela mesma.

Janet Echelman, "Filtro de sonhos" (2001)

Imagine-se num quarto vazio com paredes brancas, você já consegue se expressar ali de alguma maneira; agora pense que te deram um lápis… agora pense que te deram um pincel e uma tinta… agora pense que te deram uma câmera, uma filmadora… pense que te deram um aparelho de som…pense que te deram um projetor… Entende como as possibilidades seguem os avanços tecnológicos?

Uma coisa muito interessante também é ver como que, por vezes, os artistas não conseguem concretizar uma arte que eles idealizaram sozinhos, e precisam de suporte de terceiros para conquistar esse fim, algo bem presente na Arte Contemporânea.

O que vemos abaixo é um exemplo disso, uma instalação localizada no Museu do Inhotim, MG, Brasil. Onde a artista Janet Cardiff teve suporte de um engenheiro de som para construir essa sala acústica, onde alto-falantes posicionados estrategicamente reproduzem a voz de cada integrante de um coro, que canta “Spem in Alium”, do inglês Thomas Tallis.

Cada alto-falante contendo a voz de 1 única pessoa, reproduz a ideia de estarmos mesmo na presença de todo o coro, ainda mais quando nos aproximamos de um determinado canto desta sala, instalação a qual já visitei e recomendo fortemente.

Janet Cardiff, "Forty Party Motet: (A reworking of 'Spem in Alium' by Thomas Tallis 1573)" (2001)

E, por fim, não poderia deixar de mencionar o impacto da tecnologia no estudo da arte, que provavelmente você esteja utilizando agora. Ela é capaz de nos mostrar a qualquer momento uma obra de arte de nossa escolha em detalhes e cores. Onde, não muito tempo atrás, isso era algo exclusivo de livros que muitas vezes continham representações artísticas somente em preto e branco, e as pessoas tinham que imaginar as cores.

E esses foram alguns dos exemplos nos quais me baseei para dissertar um pouco sobre como a tecnologia surge e contribui para a arte em diversos sentidos. Poderia dizer até um pouco mais, citando exemplos mais recentes ainda, como as obras de projeção mapeada, usando projeções em esculturas, a fim de simular seu movimento; artes em realidade virtual, como as da artista Marina Abramovic; obras de arte interativas, imersivas e muito mais.

Para mim, é muito interessante conhecer tudo isso e estar presente neste mundo de possibilidades. Porque tudo mostra o quanto a tecnologia não é só aquilo que usamos e debatemos enquanto programadores, quando vemos ali um novo framework inovador ou uma solução mais otimizada para nossa necessidade específica. Mas sim, como ela é versátil e capaz também de promover evoluções a níveis tanto individuais como coletivos, e o faz muito bem.

An optimistic look at AI and our relationship with it

Gabriel Fernandes — Fri, 23 Feb 2024 15:00:00 +0000

Today, just over a year since we were introduced to the world of AI, it's amazing to see how much this technology has evolved since then.

Mom and child reading, generated by Midjourney

It's really impressive how much something so recent makes us question reality, whether what we see is real or not.

Something that cinema also motivated in its time, given the accounts of people running out of the rooms imagining that the train depicted would come out of the screen and advance on them.

Lumière Brothers (1895) Short ‘Arrival of a Train’

Art, at least Western, for a long time was obsessed with reality, as artists studied for years to make portraits increasingly faithful to what was seen.

And this was the role of art until the advent of photography, which in turn took over any incumbency about portraying reality. And art initially had a certain prejudice with photography, but continued representing not so much realism, but now emotions and expressions, as in cubism or impressionism.

Self-portrait (1907) by Pablo Picasso.

Photography itself shows us some interesting aspects, such as, for example, a photo is nothing more than dealing with chance and setting up a composition with what is given or arranged in front of those lenses at that certain angle. Where even we take several pictures in order to choose the best one among them later.

Parallels can be easily drawn between photography and AI image generation, because generating images using any model we also deal with chance often, when it surprises us with random elements which we have to specify their absence or not. And of course, we generate several to choose the best one among them.

Both final products aim to convey a message, whatever it may be, and all necessary means will be called upon to achieve a certain end.

But one thing that differs the two, and which is one of the highest points in my opinion, is:

The ability that AI has to shape something that before we could only imagine. Where, in a way, it is expanding our imagination and also gradually modifying our view of reality.

It's interesting to see how AI has been used in the corporate world, where in sectors that we once thought were creative and therefore without the possibility of the machine having any space, we already see processes that can and are being automated and replaced. Which means we will have more time for other tasks.

Robot working on computers, generated by Midjourney

But what tasks would those be?

Ok, AI can automate things, but what comes before this process, what comes after? We have personality, everything that is exchanged between human beings.

The most advanced AI can very well write a code, but who will say how the code should be? Who will dictate the necessary parameters and references?

When we go to the dentist we don't get there and ask him to take out the tooth that hurts, but rather, we get there, show the tooth, and he analyzes and tells us what we need.

A good architect helps the client understand the best choice among alternatives, and sometimes shows the solution to problems the client didn't even know they had.

A while ago I remember a certain meme circulating on the internet

Tweet on X, 2024

And I fully agree with it, not because customers are ignorant and don't know what they want, but because the best solution will always come from the one who knows the problem best and has enough background to say exactly what is needed, and what is not, according to their experiences. The most important thing will be this empathic connection and the support offered to those who need it.

How AI can free us from certain frustrations

These days I remember asking what a team partner was doing to make conversation and he told me he was doing such a service in RUST, at the time I was already surprised because he had never mentioned that he knew this language to me and was already building a complete solution in it. Then he showed me how he was learning, how since he knew a lot about JS and Node he wrote what he wanted in AI and asked it to draw parallels with the language he dominated.

I thought that was brilliant, because if we didn't have AI he would probably have to spend a much longer time on the learning curve, sometimes from scratch in RUST and take a much longer time to do what he wanted with the risk of getting demotivated and frustrated along the way.

So for the first time in the world we have the opportunity to learn a certain thing, in a personalized way, aligned with our individual method of learning.

Where else would a person who, "only learns" something if it is explained to her in the way of speaking of Master Yoda, learn so easily to build an API in python for example??

GPT request, 2024

In all technology adopted en masse by society, as in the case of Social Networks, we see their clearly both positive and negative impacts. If on one hand they aided access to misinformation, on the other they gave space to many narratives previously silenced by hegemonic media.

And as today we live in a society that frequently revisits its moral values which are constantly changing, it is very important that we understand the power of what we are generating through these AI tools, as these results can often be biased, since they are responses from a certain database, which can also end up being influenced by its products.

My perspective for the future does not try to be prophetic in any way, saying which sectors will be decimated by AI if any, but rather having to follow a path of knowing what we can do from then on, where I believe that developing a critical sense was not and will not be something in vain; because that way we will have more property to write, choose references, parameters and judge as many times as necessary the result of our requests.

In the end, AI's will be what society wants them to be.

Thinking about that famous phrase:

"Talent is hitting a target that no one hits, Genius is hitting a target that no one had seen"

How many new targets are exposed by the AI's flashlight, since that today, more than ever, we are able to easily expand our intellectual horizons?

I had the idea to write this after spending a few weeks actively seeking out and consuming a lot of content on the subject through publications and podcasts. To better internalize what I was absorbing, improve my writing, and expand my vocabulary, I decided to return to writing, but this time in a more free-form manner than when I used to post on LinkedIn, which had character limits and so on...

Feel free to comment below with any insights you may have had while reading this article. It will be a pleasure to read each one!

Um olhar otimista sobre as IA's e nossa relação com elas

Gabriel Fernandes — Fri, 23 Feb 2024 14:59:53 +0000

Hoje, pouco mais de 1 ano que fomos introduzidos ao mundo das IA's, é impressionante ver o tanto que essa tecnologia evoluiu desde então.

"Mãe e filho lendo", (2024) gerado pelo Midjourney

É realmente impressionante o quanto algo que é tão recente consegue desde já nos fazer questionar a realidade, se aquilo que vemos é real ou não.

Algo que o cinema também motivou em sua época, dados os relatos de pessoas saindo correndo das salas imaginando que o trem representado sairia da tela e avançaria sobre elas.

Irmãos Lumière (1907) Curta ‘Arrival of a Train’(Chegada do Trem)

A arte, ao menos ocidental, por muito tempo ficou obcecada com a realidade, uma vez que artistas estudavam durante anos para realizarem retratos cada vez mais fiéis ao que se era visto.

E esse foi o papel da arte até o surgimento da fotografia, que por sua vez tomou conta de qualquer incumbência sobre retratar a realidade. E a arte teve um certo preconceito inicial com a fotografia, mas seguiu representando não mais tanto o realismo, mas agora as emoções e expressões, como no cubismo ou impressionismo.

Autorretrato (1907) de Pablo Picasso.

A fotografia por si só nos mostra alguns aspectos interessantes, como por exemplo, uma foto nada mais é do que lidar com o acaso e montar uma composição com o que é nos dado ou disposto na frente daquelas lentes por aquele determinado ângulo. Onde até mesmo tiramos várias fotod a fim de escolher a melhor dentre elas futuramente.

Paralelos podem ser facilmente traçados entre a fotografia e a geração de imagens por IA, pois gerando imagens usando um modelo qualquer nós lidamos tambem com o acaso muita das vezes, quando ela nos surpreende com elementos aleatórios os quais temos que especificar sua ausência ou não. E claro, geramos várias para dentre elas escolher a que melhor se aplica.

Ambos produtos finais têm o objetivo de passar uma mensagem, seja ela qual for, e todos os meios necessários serão convocados para se conquistar determinado fim.

Mas uma coisa que difere as duas, e que é um dos pontos mais altos na minha opinião, é:

A capacidade que a IA tem de dar forma a algo que antes só poderíamos imaginar. Onde, de certa maneira, ela está ampliando nossa imaginação e também modificando gradualmente nossa visão sobre a realidade.

É interessante ver como a IA tem sido utilizada no meio corporativo, onde em setores que antes julgávamos como criativos e portanto sem a possibilidade da máquina ter algum espaço, já vemos processos que podem e estão sendo automatizados e substituídos. O que significa que teremos mais tempo para outras tarefas.

"Robo trabalhando com máquinas", (2024) gerado pelo Midjourney

Mas quais tarefas seriam essas?

Ok, a IA consegue automatizar as coisas, mas o que vem antes desse processo, o que vem depois? Temos ai pessoalidade, tudo que é troca entre seres humanos.

A IA mais avançada consegue muito bem escrever um código, mas quem dirá como o código deverá ser? Quem ditará os parâmetros e referências necessárias?

Quando vamos ao dentista não chegamos lá e pedimos para ele tirar o dente que dói, mas sim, chegamos lá, mostramos o dente, e ele analisa e fala o que precisamos.

Um bom arquiteto ajuda o cliente a entender a melhor escolha dentre as alternativas, e algumas vezes mostrar a solução para problemas os quais o cliente nem sabia que tinha.

Há um tempo atrás lembro de um certo meme circular pela internet

Tweet no X (2024)

E eu concordo plenamente com ele, não porque os clientes sejam ignorantes e não sabem dizer o que querem, mas que a melhor solução sempre virá daquele que melhor conhece o problema e tem bagagem suficiente para dizer exatamente o que é necessário, e o que não é, segundo suas experiências. O mais importante será essa conexão empática e o suporte oferecido para aqueles que precisam dela.

Como a IA pode nos livrar de certas frustrações

Esses dias lembro de perguntar o que um parceiro de time estava fazendo para puxar assunto e ele me disse que estava fazendo tal serviço em RUST, na hora eu já estranhei porque ele nunca havia mencionado que sabia essa linguagem para mim e já estava construindo uma solução completa nela. Daí ele me mostrou como estava fazendo para aprender, como ele sabia bastante de JS e Node ele escrevia o que queria na IA e pedia para ela traçar paralelos com a linguagem que ele dominava.

Achei aquilo genial, porque caso não tivéssemos a IA ele provavelmente teria de passar um tempo bem maior na curva de aprendizagem, às vezes do zero em RUST e levar um tempo bem maior para fazer o que queria com o risco de ficar desmotivado e frustrado durante o caminho.

Então pela primeira vez no mundo temos a oportunidade de aprender determinada coisa, de maneira personalizada, alinhada com nosso método individual de aprendizado.

Onde mais uma pessoa que, "só aprende" algo se isso for explicado para ela na maneira de falar do Mestre Yoda, aprenderia com tanta facilidade a construir uma API em python por exemplo??

Requisicão para o ChatGPT (2024)

Em toda tecnologia adotada em massa pela sociedade, como no caso das Redes Sociais, vemos seus claramente seus impactos tanto positivos quanto negativos. Se por um lado elas auxiliaram o acesso à desinformação, por outro ela deu espaço para muitas narrativas antes silenciadas pelos meios de comunicação hegemônicos.

E como hoje vivemos numa sociedade que revisita frequentemente seus valores morais que vivem em constante mutação, é muito importante que nós entendamos o poder do que estamos gerando através dessas ferramentas de IA, pois esses resultados por muitas vezes podem ser enviesados, uma vez que são respostas de uma determinada base de dados, que também pode acabar sendo influenciada pelos seus produtos.

Minha perspectiva para o futuro não tenta ser de maneira alguma profética, dizendo quais os setores que serão dizimados pela IA se é que serão, mas sim tendo a seguir um caminho de saber o que podemos fazer desde então, onde acredito que desenvolver um senso crítico não foi e nem será algo em vão; porque assim teremos mais propriedade para escrever, escolher referências, parâmetros e julgar quantas vezes forem necessárias o resultado de nossas requisições.

No final das contas as IA's serão o que a sociedade quiser que ela seja.

Pensando naquela famosa frase:

"Talento é acertar um alvo que ninguém acerta, Genialidade é acertar um alvo que ninguém tinha visto"

Quantos novos alvos são expostos pela lanterna da IA, já que hoje, mais do que nunca, conseguimos expandir facilmente os nossos horizontes intelectuais?

Tive a ideia de escrever isso após algumas semanas buscando e consumindo bastante do assunto em publicações e podcasts. E para fixar melhor o que estava absorvendo, melhorar a escrita e o vocabulário resolvi voltar a escrever, agora numa maneira um pouco mais livre do que quando fazia em postagens no Linkedin que tinham limite de caracteres e etc...

Sinta-se a vontade para comentar aqui embaixo qualquer insight que você possa ter tido lendo esse artigo, sera um prazer ler cada um!