
software developer @factorialEs I tweet random rants about software and politics
maputo
Joined April 2011
- Tweets9.5K
- Following710
- Followers635
- Likes17.7K
Osvaldo Maria 🇲🇿 retweeted
Conseguimos🔥🔥🔥
Hoje conseguimos identificar a classe do erro cometido pelo aluno durante sua jornada de estudo e comunicamos ao professor as falhas e o que este pode fazer para corrigi-los na aula seguinte, melhorando assim o desempenho académico daquele aluno em particular.
Osvaldo Maria 🇲🇿 retweeted
Replying to @Cheque_Senpai @impurefunction
E falando na lei da Família, tens que ajudar teus sobrinhos caso tua irmã e/ou irmão não consegue prover para eles. Mesmo em caso de pensão de alimentos e tal...🤣🤣🤣
Osvaldo Maria 🇲🇿 retweeted
Nigerian billionaire Aliko Dangote launched a $2.5 billion fuel storage and regional pipeline project in Namibia's Walvis Bay on Wednesday that could reshape the fuel supply architecture in Southern Africa. reuters.com/world/africa/bil…
Osvaldo Maria 🇲🇿 retweeted
If you want to grow in a field, you really need to filter what you choose to listen to.
You’ll hear a lot about how hard it is to get a job, how competitive the industry is, and how difficult it is to break in.
It’s good to know the reality, but don’t let those conversations kill your zeal before you even get started.
Be informed, but don’t let fear stop you from trying. Be intentional about what you feed your mind. You need the right information, not constant discouragement.
Osvaldo Maria 🇲🇿 retweeted
Today we're coming out of stealth, introducing Neywa Labs 🚀
We built the infrastructure to build and deploy AI that runs on the devices people already own. Private, offline, unlimited.
One mission: make AI available to everyone in the world.
neywalabs.ai/
Osvaldo Maria 🇲🇿 retweeted
Oi, Boa tarde à todos!
Meu nome é Ivo Frederico Augusto Gundanhane tenho 25 anos.
Sou técnico médio de sistemas electromecânicos e estou em busca de uma oportunidade de emprego, tenho 3 anos de experiência, na área de manutenção electrica.
the right way to use model capabilities is not to ship 10x more features to prod
it's to spend more time understanding your users, trying experiments, building prototypes, learning about things you don't understand so that you can ship things that actually work
Osvaldo Maria 🇲🇿 retweeted
Beware people obsessed with outcomes instead of building outcome machines. Its worse than ever with AI, but these people existed before. Short term results above all else, etc. Don't fall into the trap. Invest in building strong fundamentals, invisible supports, and outcomes flow like water. An outcome machine.
Osvaldo Maria 🇲🇿 retweeted
Agent-centric systems assume lessons belong inside the agent.
But forcing a single agent to absorb every update creates redundant, task-specific, and conflicting lessons that dilute performance. Each new adaptation risks breaking what already worked.
Externalizing intelligence breaks the single-agent bottleneck. Knowledge improves, while agents become generic, stateless, and disposable. The result is a living artifact of insights—fully inspectable, versionable, and transferable to any model or human.
Explore the knowledge base here: recursive-knowledge.github.i…
Osvaldo Maria 🇲🇿 retweeted
Huge respect to Italy who took this stance a decade ago...
All European nations to boycott World Cup if Fifa goes ahead with controversial investment plan bbc.in/3RDovXM
Osvaldo Maria 🇲🇿 retweeted
can a group of goated designers please take lsd together to unlock their inner steve jobs and cook a new ai interaction paradigm that's not chat
Osvaldo Maria 🇲🇿 retweeted
Afinal voces pagam curso de cartão virtual? Moz ya🤣
Redot pay
Zypto pay
Meru-use airtm to receive from mpesa
You’re welcome. Follow instructions 🙏🏽aceita mpesa e bank transfer
Osvaldo Maria 🇲🇿 retweeted
A partir de hoje o software de gestão @BuinozBZ passou a correr inteiramente no Google Cloud (GCP). Saímos da AWS em Ohio e movemos tudo para o GCP em Joanesburgo.
O resultado vai além da velocidade, que ficou cerca de 87% mais rápida, os custos irão baixar em cerca de 60%.
A nossa infra estava na AWS, em Ohio. Um EC2 a correr o servidor de aplicação, um RDS para a base de dados. Cada chamada da app em Moçambique atravessava o Atlântico, batia em Ohio e voltava. Mais de 300 milissegundos só de viagem. Numa página com cinco pedidos, perdíamos quase dois segundos só em rede.
Hoje está tudo no GCP em Joanesburgo: Cloud Run no lugar do EC2. Cloud SQL no lugar do RDS. A mesma viagem demora 40 milissegundos. Sem mexer numa linha de código.
A app continua a ser desenvolvida em @FlutterDev e Hasura. Não mexemos. O front-end nunca foi o problema. O desafio era afastar o back-end de Ohio e trazê-lo para Joanesburgo, perto de quem o usa.
Em 2024 a Google abriu a região em Joanesburgo. Quisemos lá estar assim que fizesse sentido para os nossos utilizadores. Não é missão. É geografia a começar a fazer sentido.
Para o utilizador é só a app mais rápida. Joanesburgo hoje. Maputo amanhã, quando chegar o Datacenter. Para já: rápido e próximo.
Osvaldo Maria 🇲🇿 retweeted
Okay, vocês dizem que há combustível suficiente no país, já subiram de preço há uns dias… qual é problema agora ? Porque temos que estar em filas longas ainda ? Qual é a dificuldade agora ? Está ser exaustivo.
Osvaldo Maria 🇲🇿 retweeted
Engineering is not only about finding the right abstractions but also the right constraints. What appears easy today is often the result of decades spent distilling complexity. This process is often unfortunately unrecognized.
Osvaldo Maria 🇲🇿 retweeted
Sobre VM7: é um facto, o homem está a fazer oposição a sério!
Democracia não é apenas campanha eleitoral e votação.
Osvaldo Maria 🇲🇿 retweeted
OpenClaw - the agentic software spreading like wildfire - was built on top of Pi, a minimalist, self-modifying agent. I sat down with Pi's creator, @badlogicgames and longtime Pi user (+ the creator of Flask) @mitsuhiko to talk Pi, and their (very grounded!) takes on building with AI.
Timestamps:
00:00 Intro
07:30 How Mario, Armin, and Peter Steinberger met
15:15 How 30 dev teams use AI agents: learnings
21:50 The importance of judgment
24:26 Challenges when non-engineers write code
28:30 Downsides of over-automation
32:18 Pi
48:09 OpenClaw + Pi
50:54 “Clankers”
57:32 Open source and AI
1:00:22 Complexity as the enemy
1:02:50 Building an AI-native startup
1:11:52 “Slow the F down”
1:16:40 MCPs vs. CLI
1:25:03 Predictions and staying up to date
• YouTube: youtu.be/n5f51gtuGHE
• Spotify: open.spotify.com/episode/1fD…
• Apple: podcasts.apple.com/us/podcas…
Brought to you by:
• @statsig – The unified platform for flags, analytics, experiments, and more. statsig.com/pragmatic
• @SonarSource — The makers of SonarQube, the industry standard for code verification and automated code review. Try it out for yourself. sonarsource.com/plans-and-pr…
• @WorkOS – WorkOS gives you APIs to ship enterprise features – SSO, directory sync, RBAC, audit logs – in days, not months. Visit WorkOS.com to learn more.
---
Three parts I found especially interesting in this discussion:
1. New trend: AI makes it harder for senior engineers to reject pointless complexity.
Historically, senior engineers kept software complexity at bay simply by saying “no” a lot. But Armin observes that these days, more junior engineers and product managers deploy agent-scripted counterarguments when a senior colleague kicks an idea to the curb. This makes decision-making exhausting, and more bad ideas make it into production as a result.
2. It should be MUCH easier to build specialized tools for specific tasks.
Different projects need different harness types because, as Mario points out, the same hammer is not ideal for every single construction job. As such, Pi is built with the goal of allowing the creation of specialized harnesses. It can modify itself so that a user can create the bespoke harness needed for any task. Mario believes it’s a preview of how self-modifiable software might look in the future.
3. Automation bias is one of the biggest risks of working with AI agents.
Once devs confirm that an AI agent can produce acceptable code, they start to review its output less often, even though agents can – and do! – produce slop. Mario advises being far more sceptical with agents, and cautions that the quality of their output isn’t guaranteed, however well they performed previously.
Osvaldo Maria 🇲🇿 retweeted
Researchers sent the same resume to an AI hiring tool twice. Same qualifications. Same experience. Same skills. One version was written by a real human. The other was rewritten by ChatGPT.
The AI picked the ChatGPT version 97.6% of the time.
A team from the University of Maryland, the National University of Singapore, and Ohio State just published the receipt. They took 2,245 real human-written resumes pulled from a professional resume site from before ChatGPT existed, so the human writing was actually human. Then they had seven of the most-used AI models in the world rewrite each one. GPT-4o. GPT-4o-mini. GPT-4-turbo. LLaMA 3.3-70B. Qwen 2.5-72B. DeepSeek-V3. Mistral-7B.
Then they asked each AI to pick the better resume. Every model picked itself.
GPT-4o hit 97.6%. LLaMA-3.3-70B hit 96.3%. Qwen-2.5-72B hit 95.9%. DeepSeek-V3 hit 95.5%. The real human almost never won.
Then the researchers tried the obvious objection. Maybe the AI is just better at writing. So they had real humans grade the resumes for actual quality and ran the experiment again, controlling for it. The result was worse. Each AI kept picking itself even when human judges rated the human-written version as clearer, more coherent, and more effective.
It gets worse. The AIs do not just prefer AI over humans. They prefer themselves over other AIs. DeepSeek-V3 picked its own resumes 69% more often than LLaMA's. GPT-4o picked its own 45% more often than LLaMA's. Each model can recognize and reward its own dialect.
Then the researchers ran the simulation that ends careers. Same job. 24 occupations. Same qualifications. The only variable was whether the candidate used the same AI as the screening tool. Candidates using that AI were 23% to 60% more likely to be shortlisted. Worst gap was in sales, accounting, and finance.
99% of large companies now run AI on incoming resumes. Most of them use GPT-4o. The paper just proved GPT-4o picks GPT-4o 97.6% of the time.
If you wrote your own cover letter this week, you did not lose to a better candidate. You lost to a worse candidate who paid OpenAI 20 dollars.
Your qualifications do not matter if the AI prefers its own handwriting over yours.
























