Published:
AI tools have made certain parts of programming faster in ways that are hard to argue with. Generating a function from a description, explaining what an error message means, writing test cases for a known interface, translating code between languages — these tasks take a fraction of the time they did two years ago. That efficiency gain is real and significant. Read more
Published:
Google released Gemini 3 on November 18, 2025, making it available the same day in the Gemini app (replacing Gemini 2.0 Flash as the default), Google AI Studio, Vertex AI, and Google Search’s AI Mode — the generative answer panel that had been gradually expanding across Search since mid-2024. Gemini 3 came in Ultra, Pro, and Flash tiers; the Flash variant was designed for low-latency applications at lower per-token cost, while Ultra targeted complex reasoning and multimodal tasks. Read more
Published:
Every November 10, Turkiye remembers Mustafa Kemal Ataturk. Read more
Published:
Ubuntu 25.10 “Questing Quokka” was released on October 23, 2025, the interim release immediately preceding Ubuntu 26.04 LTS, scheduled for April 2026 with five years of standard enterprise support. Like all non-LTS Ubuntu releases, 25.10 received only nine months of security maintenance (through July 2026), targeting developers, enthusiasts, and infrastructure teams who want current software versions rather than the stability-optimized LTS freeze. The release shipped GNOME 48, which introduced a tiled window management feature (GNOME’s response to years of community requests for built-in tiling without extensions), updated color management, and improved notification grouping. The included Linux kernel was 6.12, which had reached upstream LTS status in December 2024, adding improved scheduler latency for interactive workloads and continued refinement of the SCHED_EXT extensible scheduler framework that allows user-space BPF programs to implement custom scheduling policies. Read more
Published:
Welcome to Computer Engineering in 2025. Read more
Published:
A programming assignment looks simple to describe: write a program that does something, test it, submit it. But the difference between an assignment that produces genuine learning and one that produces submitted code is significant, and it is mostly in the design. Read more
Published:
Apple announced the iPhone 17, iPhone 17 Plus, iPhone 17 Pro, and iPhone 17 Pro Max on September 9, 2025, with sales beginning September 19. The standard iPhone 17 received the most significant redesign in years: a slimmer aluminum-and-glass form factor and an A18 chip (previously exclusive to Pro models in 2024), bringing faster Neural Engine performance and improved camera processing to the base lineup. iPhone 17 Air — replacing the Plus as the thin-and-light option — launched at 5.5mm thin, the slimmest iPhone ever, at the cost of a smaller battery. Read more
Published:
In 2025, computers can already generate code, explain errors, summarise documents, create images, and assist with many tasks that once required substantial manual work. It is reasonable for a student to ask: if AI can do more of the work, why study Computer Engineering? Read more
Published:
I like clever algorithms, but I often enjoy simple algorithms more. There is something satisfying about a solution that looks almost obvious after you understand it — where the entire idea fits in your head at once. Read more
Published:
OpenAI released GPT-5 on August 7, 2025, available to ChatGPT users across Free, Plus ($20/month), and Pro ($200/month) tiers and via the API. Rather than launching as a separate model requiring users to switch manually from GPT-4o or o3, GPT-5 was positioned as a unified system: it handled conversational and creative tasks with the low latency of earlier models while applying extended reasoning — essentially the o-series chain-of-thought approach — automatically when the system judged a problem hard enough to warrant it. Read more
Published:
OpenAI launched ChatGPT Agent in July 2025, combining GPT-5’s language understanding with a set of tools — web browser control, Python code execution via a sandboxed interpreter, file reading and writing, and external API calls — into a product designed to complete multi-step tasks with minimal user prompting per step. Unlike the earlier Operator (launched January 2025 for browser-only tasks) or Code Interpreter (launched June 2023 for data analysis in a stateless sandbox), ChatGPT Agent maintained state across a full session: it could browse the web, download a CSV file, write a Python script to process it, debug an error in the script, generate a chart, and return the result in one continuous workflow without the user re-explaining context between steps. Read more
Published:
If I were building a gaming PC on July 1, 2025, this is the machine I would actually want. I am not simply choosing the most expensive part in every category. The priority is gaming performance, then reliability, cooling and enough headroom to keep the system useful for years. Read more
Published:
Some of my favorite personal projects started with a thought that is hard to justify: I want to see what happens if I make this. No one asked for it. No deadline. No user waiting. Sometimes no clear reason beyond the fact that the idea kept returning to my mind until I decided to try it. Read more
Published:
Apple’s WWDC 2025 (June 9–13, Cupertino) introduced iOS 19, iPadOS 19, and macOS 16, with a significant portion of the developer keynote dedicated to expanding the Foundation Models framework that had debuted in iOS 18/macOS 15 (shipped September 2024). The Foundation Models framework allowed third-party applications to call Apple’s on-device approximately-3-billion-parameter language model without packaging model weights in the app bundle or routing requests to external APIs. iOS 19’s expansion added a Streaming API (returning partial tokens as they were generated rather than waiting for the full response), a multi-modal input API (accepting image and text prompts for the same model), and task-specific Adapter APIs — lightweight fine-tuned adapter layers (roughly 100 MB each) that developers could download and attach to the system model to specialize it for domain-specific tasks like legal document analysis or code review without replacing the base model. Read more
Published:
Programming looks very logical from the outside. Teaching programming is much more emotional. A student can understand a difficult idea in five minutes and then spend half an hour looking for a missing semicolon. Another student may struggle with the same topic for days and suddenly understand it after one small example — the third or fourth version of an explanation you were not sure was any different from the others. Read more
Published:
Google I/O 2025 was held May 20–21, 2025, at the Shoreline Amphitheatre in Mountain View, and the two-day developer conference was organized almost entirely around Gemini integration across Google’s product surface. The headlining model announcement was Gemini 2.5 Pro, which Google described as its most capable model to date, alongside Gemini 2.5 Flash for high-throughput lower-cost applications. AI Mode in Google Search — an opt-in interface replacing the standard ten-blue-links results page for complex queries — was expanded from its experimental state to a broader US rollout, presenting multi-step AI-generated responses with citations and follow-up query suggestions. Project Astra, Google’s universal AI assistant research prototype using a smartphone camera and screen as real-time context (demonstrated internally in 2024), was announced as the foundation for Gemini Live’s expanded capabilities: the live version of Gemini on Android could now see and respond to what the phone’s camera was pointed at, maintaining a spoken conversation while analyzing video frames in real time. Jules, an autonomous coding agent integrated with GitHub, was announced as capable of reading a GitHub issue, writing a fix across multiple files, running tests, and opening a pull request — competing directly with Anthropic’s Claude Code and GitHub Copilot Workspace in the agentic developer tooling space. Read more
Published:
AI can write code very quickly now. It can produce a function, explain an error, or suggest a solution in seconds. Because of this, I sometimes hear a simple question: if AI can write the code, do students still need to learn data structures? Read more
Published:
Meta released Llama 4 Scout and Maverick on April 5, 2025, marking the Llama family’s shift from dense transformers to mixture-of-experts (MoE) architecture. Scout had 17 billion active parameters out of 109 billion total (16 experts, one selected per token), and a 10 million token context window — the longest of any openly released model at the time. Maverick had 17 billion active parameters out of 400 billion total (128 experts) and achieved competitive scores against GPT-4o and Gemini 2.0 Flash on coding and reasoning benchmarks. A third model, Llama 4 Behemoth at over 2 trillion total parameters, was announced as still in training. Read more
Published:
NVIDIA held GTC 2025 in San Jose from March 17–21, with Jensen Huang’s keynote detailing the Blackwell Ultra accelerator and announcing the Rubin architecture roadmap. Blackwell Ultra — the B200 Ultra — raised HBM3e memory to 288 GB per GPU (up from 192 GB on the B200) and increased FP4 throughput to support the larger context windows and reasoning chains frontier models were beginning to require. NVLink 6 connected up to 576 B200 Ultra GPUs within a single NVLink domain, enabling the full weight matrix of a large model to reside in fast GPU memory rather than requiring frequent host-memory transfers. Read more
Published:
OpenAI released Deep Research on February 2, 2025, initially for ChatGPT Pro subscribers ($200/month). The feature used an internal version of the o3 reasoning model to autonomously conduct multi-step research tasks: planning a search strategy, executing dozens of web searches, reading and synthesizing source documents, tracking intermediate findings, and producing a 2,000–5,000 word report with inline citations. A typical Deep Research task ran for 5–30 minutes without user intervention, producing outputs comparable in depth to what a research analyst might produce in a few hours. The model was given access to Bing web search and document reading tools; it could adapt its search plan based on what it found at each step, a form of “reflection” where the model assessed the quality of evidence and decided whether to investigate further or broaden scope. Google had launched Gemini Deep Research on December 9, 2024 (powered by Gemini 1.5 Pro with Google Search), making this the first major head-to-head competition between deep research agents within weeks of each other. Read more