In the tech world, there is a constant flow of changes and keeping up with them means the choice for tools and technologies which are the most appropriate to invest your time in.
In 2026 the best programming language or technology stack to learn really depends on your personal aims, hobbies, and apps you are going to create.
The use of AI is increasing. AI as a “Pair Programmer” is becoming the default. Code completion, refactoring, and boilerplate generation are used often. Devs spend more time reviewing and steering code than typing it. “Explain this error” and “why is this slow?” prompts are useful.
In prompt-Driven Development programmers describe the intent in natural language and then let AI generate first drafts of functions, APIs, or configs. Iterate by refining prompts rather than rewriting code. Trend: Knowing how to ask is becoming as important as syntax.
Strong growth in: Auto-generated unit and integration tests and edge-case discovery. Trend: “Test-first” is easier when AI writes the boring parts.
AI is moving up the stack. Trend: AI as a junior architect or reviewer, not the final decider.
AI comes to Security & Code Quality Scanning. Rapid adoption in: Static analysis and vulnerability detection, secret leakage and dependency risk checks. AI can give secure-by-default code suggestions. Trend: AI shifts security earlier in the SDLC (“shift left”).
Instead of one-off prompts: AI agents that plan → code → test → fix → retry. Multi-step autonomous tasks (e.g., “add feature X and update docs”) can be done in best cases. Trend: Still supervised, but moving toward semi-autonomous dev loops.
AI is heavily used for explaining large, unfamiliar codebases and translating between languages/frameworks. It helps onboarding new engineers faster.
What’s changing: Less manual boilerplate work
More focus on problem definition, review, and decision-making. There is stronger emphasis on fundamentals, architecture, and domain knowledge. Trend: Devs become editors, designers, and orchestrators.
AI usage policies and audit trails is necessary. Trend: “Use AI, but safely.”
Likely directions:
Deeper IDE + CI/CD integration
AI maintaining legacy systems
Natural-language → production-ready features
AI copilots customized to your codebase
561 Comments
Tomi Engdahl says:
Wine 11 rewrites how Linux runs Windows games at the kernel level, and the speed gains are massive
https://www.xda-developers.com/wine-11-rewrites-linux-runs-windows-games-speed-gains/
Tomi Engdahl says:
Why High-Performance Rust Rarely Looks “Idiomatic Rust”
The first time I felt this properly, it hurt my ego more than my code.
https://medium.com/@build_break_learn/why-high-performance-rust-rarely-looks-idiomatic-rust-cd14204ed20a
Tomi Engdahl says:
Harness design for long-running application development
Published Mar 24, 2026
Harness design is key to performance at the frontier of agentic coding. Here’s how we pushed Claude further in frontend design and long-running autonomous software engineering.
https://www.anthropic.com/engineering/harness-design-long-running-apps
Tomi Engdahl says:
https://javarevisited.blogspot.com/2026/03/top-9-system-design-and-software-design-books-and-resources.html
Tomi Engdahl says:
Vibe Coding a Private AI Financial Analyst with Python and Local LLMs
Learn to build an AI data analyst with Python: analyzes data, detects anomalies, and generates predictions using local LLMs.
https://www.kdnuggets.com/vibe-coding-a-private-ai-financial-analyst-with-python-and-local-llms
Tomi Engdahl says:
https://blog.jetbrains.com/blog/2026/03/24/introducing-jetbrains-central-an-open-system-for-agentic-software-development/
Tomi Engdahl says:
https://github.com/RustedBytes/rsloop
Tomi Engdahl says:
TypeScript Just Became the #1 Language on GitHub by Raw Usage. JavaScript Developers Are Having an Identity Crisis.
https://medium.com/@coders.stop/typescript-just-became-the-1-language-on-github-by-raw-usage-c51962f3c575
Tomi Engdahl says:
CodingPrep
Open Source coding interview prep tool with AI interviewer
https://www.producthunt.com/products/codingprep
Tomi Engdahl says:
https://www.theinformation.com/newsletters/ai-agenda/claude-gaining-openclaw
Tomi Engdahl says:
I fine-tuned a 7B model to write my Home Assistant automations, and it actually works
https://www.xda-developers.com/fine-tuned-7b-model-home-assistant-automations/
Tomi Engdahl says:
Will AI Replace Software Developers?
#
ai
#
programming
#
discuss
#
news
Lately, the question “Will AI replace us?” has worried many people. We can see how LLMs handle programming tasks very well and write code at a middle to senior level. This makes many software developers concerned about their future.
https://dev.to/empiree/will-ai-replace-software-developers-1fo0
Tomi Engdahl says:
When AI turns software development inside-out: 170% throughput at 80% headcount
https://venturebeat.com/orchestration/when-ai-turns-software-development-inside-out-170-throughput-at-80-headcount
Tomi Engdahl says:
https://tekai.fi/
Tomi Engdahl says:
CIOs reimagine software’s future as AI agents advance
Feature
Mar 27, 2026
6 mins
SaaS isn’t dying any more than on-premises software, but IT leaders face huge challenges managing all manner of AI as it seeps deeper into the systems that fuel business operations.
https://www.cio.com/article/4148303/cios-rethink-softwares-future-as-ai-agents-advance.html
Tomi Engdahl says:
https://linuxiac.com/vitruvianos-0-3-debuts-as-haiku-inspired-linux-os/
Tomi Engdahl says:
https://www.brentozar.com/archive/2026/03/using-claude-code-with-sql-server-and-azure-sql-db/
Tomi Engdahl says:
https://www.infoworld.com/article/4141944/how-generative-ui-cut-our-development-time-from-months-to-weeks.html
Tomi Engdahl says:
https://blog.stackademic.com/the-lazy-way-to-become-a-programmer-is-actually-the-smart-way-524b06ae8fe7
Tomi Engdahl says:
https://www.xda-developers.com/ran-ollama-open-webui-on-200-mini-pc-local-ai-stack-actually-works/
Tomi Engdahl says:
https://techcrunch.com/2025/04/29/microsoft-ceo-says-up-to-30-of-the-companys-code-was-written-by-ai/
Microsoft CEO says up to 30% of the company’s code was written by AI
During a fireside chat with Meta CEO Mark Zuckerberg at Meta’s LlamaCon conference on Tuesday, Microsoft CEO Satya Nadella said that 20% to 30% of code inside the company’s repositories was “written by software” — meaning AI.
Nadella gave the figure after Zuckerberg asked roughly how much of Microsoft’s code is AI generated today. The Microsoft CEO said the company was seeing mixed results in AI-generated code across different languages, with more progress in Python and less in C++.
Microsoft CTO Kevin Scott previously said he expects 95% of all code to be AI generated by 2030.
When Nadella threw the question back at Zuckerberg, the Meta CEO said he didn’t know how much of Meta’s code is being generated by AI.
On Microsoft rival Google’s earnings call last week, CEO Sundar Pichai said AI was generating more than 30% of the company’s code. Of course, it’s unclear how exactly Microsoft and Google are measuring what’s AI generated versus not, so these figures are best taken with a grain of salt.
Tomi Engdahl says:
https://sii.pl/en/events/multiagent-ai-for-embedded-software-development/
Tomi Engdahl says:
https://github.com/openai/codex-plugin-cc
Tomi Engdahl says:
https://www.kdnuggets.com/getting-started-with-smolagents-build-your-first-code-agent-in-15-minutes
Tomi Engdahl says:
https://dev.to/somadevtoo/i-found-leetcode-for-system-design-and-its-awesome-2h7f
Tomi Engdahl says:
Vercel Releases JSON-Render: a Generative UI Framework for AI-Driven Interface Composition
https://www.infoq.com/news/2026/03/vercel-json-render/
Tomi Engdahl says:
Muutkin on huomanneet saman kuin minä (siihen meni 10 minuuttia kun aloitin). AI koodaus toimii hyvin jos sen laittaa tekemään featuren kerrallaan, joka aina testataan toimivaksi.
Netflix, Meta, and IBM speakers: AI will make anyone a 10x programmer, but with 10x the cleanup
Agents to check the work of the agents
https://www.theregister.com/2026/04/04/all_things_ai_conference/
All Things AI AI is easy to use, but not quite as easy as just barking “Alexa! Make me an e-commerce site.” And, no, adding “DON’T HALLUCINATE” to the instruction loop won’t help.
More to the point, optimal AI results favor the well-fortified agent, according to speakers from IBM, Meta, and Netflix – among others – at the All Things AI conference in Durham, North Carolina.
The more you want AI to do your bidding, the more preparatory chores you’ll need to do, they advised.
Numerous talks evoked the Jevons Paradox, where the more efficient a resource becomes, the more it’s used. The paradox is often used to explain why AI won’t take everyone’s jobs. In fact, it will create more jobs, the argument goes.
Currently, AI is certainly creating more work for its users, requiring time to prepare context and check outcomes. Claude will make anyone a 10x programmer, but they’ll need to clean up 10x the results.
Or, in the most apocalyptic terms, before the singularity can enslave humankind as energy pods à la The Matrix, it will require some assistance from us meat sacks to get around.
The sorcerer’s apprentice
How is AI keeping the folks at Netflix busy? In a talk, Netflix UI architect Ben Ilegbodu explained how as soon as you create an agent to automate some task, you will need a second agent to evaluate the work done.
Ilegbodu sometimes even breaks the job into multiple agents that specialize in different parts of the code review. He calls this approach “adversarial code review.”
Oh, you’ll also need a third agent to orchestrate the actions between the first two, he said.
Ilegbodu’s workday is the Jevons Paradox incarnate. Once he sets off one agent to implement some new feature, he tasks another agent to do the preliminary work for the next task he has in mind. In effect, he is “parallelizing himself so the work is always happening.”
AI has allowed Ilegbodu to code in languages he doesn’t yet know, such as Python, Bash, and Groovy.
But this context switching can get wearisome, he admitted. “At the end of the day, I’m actually kind of tired, because effectively, I spent the whole day talking to something.”
The insatiable intern
Many coders think about AI like an eager junior developer on the team: enthusiastic but naive. But unlike a junior dev, an AI won’t “get overwhelmed,” said Meta Developer Advocate Justin Jeffress, in his talk.
You can just keep shoveling more information to the AI, and it will take it all in (for as many tokens as you can afford).
Such bottomless hunger leads to what Jeffress called “context rot.”
“Over time, as you interact with your AI agent, the more stuff it has to calculate to provide an answer, the more there is vying for its attention and the less likely it’s going to do the right thing,” he said.
Vague instructions lead to diffuse results, he told the audience. Clearly thinking about what information you are giving to the agent is the work of context engineering, which, in the short time of agentic AI, has become an art form, if not quite a proper discipline yet.
Wishful prompting
The fact that the AI doesn’t do exactly what you want it to do is not a problem with the AI. It’s a problem with your lack of “decomposition” skills, posited Luis Lastras, IBM director of language and multimodal technologies, in his talk.
Wishful prompting is just typing “I must insist, do not hallucinate. My career depends on it, please, please, please.” It’s like casting a spell and hoping it’ll work, he said.
Instead, developers should be thinking about how to break the work up into smaller, more bite-sized portions for the agent.
Tomi Engdahl says:
Open Multi-Agent
TypeScript framework for multi-agent orchestration. One runTeam() call from goal to result — the framework decomposes it into tasks, resolves dependencies, and runs agents in parallel.
https://github.com/JackChen-me/open-multi-agent/blob/main/README.md
Tomi Engdahl says:
Microsoft Agent Framework Version 1.0
https://devblogs.microsoft.com/agent-framework/microsoft-agent-framework-version-1-0/
Tomi Engdahl says:
Rust Meets Golang in a New Programming Language
Lisette is the answer to a burning question: how far can Go be evolved without losing its productivity?
https://analyticsindiamag.com/ai-features/how-one-developer-is-rethinking-go-using-rust
Tomi Engdahl says:
I Built an AI Agent Team for Software Development and Tested on 5 Real Projects
I assigned agents to PM, SWE, QA, and on-call roles and used the setup across five different software projects.
https://alexeyondata.substack.com/p/i-built-an-ai-agent-team-for-software
Tomi Engdahl says:
https://github.com/JackChen-me/open-multi-agent/blob/main/README.md
Tomi Engdahl says:
Why pgEdge thinks MCP (not an API) is the right way for AI agents to talk to databases
pgEdge launches a production-ready MCP Server for Postgres, bringing AI agent connectivity, schema introspection, and reduced token usage to any Postgres database.
https://thenewstack.io/pgedge-mcp-postgres-agents/
Tomi Engdahl says:
https://thenewstack.io/broadcom-donates-velero-cncf/
Tomi Engdahl says:
Apr 2, 2026
—
10:00 am
PT
Ship Code Faster with Claude Code on Vertex AIShip Code Faster with Claude Code on Vertex AI
Explore how to build and access Claude Code on Vertex AI to accelerate your development workflows. In this technical webinar, we’ll walk through Claude Code’s capabilities and how to set it up on Vertex AI. Through live demonstrations, you’ll see how Claude Code handles real coding tasks end to end so you can start building with confidence on Vertex AI. We’ll also answer some of the top developer community questions for building with Claude on Vertex AI.
https://www.anthropic.com/webinars/ship-code-faster-with-claude-code-on-vertex-ai
Tomi Engdahl says:
https://dev.to/sylwia-lask/9-things-youre-overengineering-the-browser-already-solved-them-o99
Tomi Engdahl says:
I ditched Claude Code and OpenCode for Pi, and my coding workflow became predictable again
https://www.xda-developers.com/replaced-claude-code-and-opencode-with-pi/
Tomi Engdahl says:
Introducing EmDash — the spiritual successor to WordPress that solves plugin security
https://blog.cloudflare.com/emdash-wordpress/
Tomi Engdahl says:
How To Vibe Code Your Entire Marketing Engine In A Day
https://www.forbes.com/sites/jodiecook/2026/03/30/how-to-vibe-code-your-entire-marketing-engine-in-a-day/
Tomi Engdahl says:
https://thenewstack.io/how-to-integrate-vs-code-with-ollama-for-local-ai-assistance/
Tomi Engdahl says:
I built a clipboard server on the Arduino Uno Q, and it replaced a workflow I didn’t realize I hated
https://www.xda-developers.com/built-clipboard-server-arduino-uno-q-replaced-workflow/
If you work across multiple devices, you’ve almost certainly emailed yourself a link, pasted something into a notes app just to open it on another machine, or maybe even typed out a URL by hand because you couldn’t be bothered to find a better way. Clipboard sync between devices is a largely solved problem by now, but sometimes it’s nice to just have a simple way to quickly and easily copy text between two devices without any additional software. Sure, Apple has Universal Clipboard, Windows has its own clipboard sync, and third-party tools like ClipCascade let you self-host something, but none of them felt right for the level of simplicity that I wanted
Tomi Engdahl says:
PEP 816: How Python is getting serious about Wasm
feature
Apr 1, 2026
7 mins
A newly approved Python Enhancement Proposal clarifies how Python will adhere to the WebAssembly standards going forward. Here’s what you need to know.
https://www.infoworld.com/article/4150052/how-python-is-getting-serious-about-wasm.html
Tomi Engdahl says:
I connected Claude Code to my home server through MCP, and now I manage my entire lab by talking to it
https://www.xda-developers.com/connected-claude-code-through-mcp-manage-entire-lab-by-talking/
Setting up a homelab server is straightforward. Install a base operating system, choose a Docker tool like Portainer, get the Compose file for a service, and deploy it. Managing the server and its services is where it gets tedious. Fixing an issue isn’t the tough part; finding the root cause takes more time. For a simple issue, I had to open Portainer to find the container, SSH to read the logs, then switch to Uptime Kuma to check whether it was actually down, and Beszel to cross-check the server usage. Even if I got these things right, I had to put the pieces together to understand what the actual problem was.
Tomi Engdahl says:
https://thenewstack.io/cursor-self-hosted-coding-agents/
Tomi Engdahl says:
https://hackaday.com/2026/03/31/how-small-can-a-linux-executable-be/
Tomi Engdahl says:
VSTCM – the v.st Colour Mod – a colour vector graphics generator
The vstcm is a vector signal transceiver PCB which generates colour vector graphics to be displayed on an oscilloscope or vector monitor such as Amplifone, Wells Gardner WG6100, Hantarex MTRV and Electrohome G05, as used in Star Wars, Tempest, Gravitar, etc.
https://github.com/english1234/vstcm?tab=readme-ov-file
Tomi Engdahl says:
https://hackaday.com/2026/03/26/modern-smartphone-vs-80s-supercomputer/
Tomi Engdahl says:
Python in Excel isn’t just for programmers—4 useful things you can do with it right now
https://www.howtogeek.com/microsoft-excel-python-isnt-just-for-programmers/
For decades, if you wanted to do heavy data lifting in Excel, your options were nested formulas, VBA macros, or a trip into the Power Query Editor. But with Python now accessible directly from Excel cells, there’s a fourth way. It lets you perform complex data cleaning, sentiment analysis, and more—all without leaving the formula bar.
While Python has traditionally lived in complex coding environments, Microsoft has now brought it directly into the Excel grid. The best part? You don’t need to be a programmer to use it. If you can copy and paste a formula and understand how an Excel table works, you’re already halfway there. It’s less about writing code from scratch and more about using ultra-powerful functions to do things faster and more flexibly than standard Excel.
Technically, Python doesn’t run on your local computer; it operates within the Microsoft Cloud. This means it won’t slow down your laptop (all the “thinking” happens on Microsoft’s high-speed servers), but you do need an internet connection to use it. The gateway to this power is the =PY() function. Typing =PY into the formula bar and pressing Tab opens Excel’s Python editing mode.
Excel’s Python runtime is powered by an Anaconda-curated Python environment that includes popular data science libraries like pandas (the workhorse for data tables) and seaborn (the expert for high-end charts). These are like “power-up” kits that give Python the ability to clean, analyze, and visualize data with minimal effort.
From a security standpoint, the process is designed to be highly controlled and “sandboxed.” Python reads your Excel tables via the xl() function, does the heavy lifting in the cloud, and then “spills” the results back into your cells. Because the environment is sandboxed, Python can’t access your local files—it only interacts with the specific data you explicitly choose to send it.
Tomi Engdahl says:
https://github.com/ultraworkers/claw-code
Tomi Engdahl says:
https://www.xda-developers.com/dlss-5-is-further-proof-that-rasterizations-days-are-numbered/