Abdella

Electronics Engineer

[Abdella @MohAbdella]

Embedded Engineering Roadmap. https://t.co/XL9NBK0xSO

[Abdella @MohAbdella]

@blind_via Wild score... must be a single layer!

[Abdella @MohAbdella]

@blind_via Bad news

[Abdella @MohAbdella]

@_luckyhada Really great work! I'd like to use the dataset but I wanted to ask you about the correctness of the schematics, how did you ensure that everything included is tested and proven to work?

[Abdella @MohAbdella]

@roman_koshchei @DavidBGrys You have pin numbers on the symbols, so you know if pins are close or not. Also cross-probing helps highlight the pin in the layout when you select it in the schematic and vice versa, you need to make sure it's turned on.

[Abdella @MohAbdella]

@abijahkaj @comma_ai Nice! How long it takes them to review and accept the design?

[Abdella @MohAbdella]

@blind_via @abijahkaj Multiple submissions are allowed

[Abdella @MohAbdella]

The historic turning point

[OpenAI @OpenAI]

We've trained an unsupervised language model that can generate coherent paragraphs and perform rudimentary reading comprehension, machine translation, question answering, and summarization — all without task-specific training: https://t.co/sY30aQM7hU

[Abdella @MohAbdella]

@ivanvnucec @blind_via @comma_ai Don't use bga. Leads to more layers and vias which is more punished than the increased PCB area

[Abdella @MohAbdella]

@NERZ_z Because for example in the highlighted area, L1 tracks cross a cut in the fill on L2. The GND fill should be continuous under the signals to maintain signal integrity. Having a fill + vias on L1 helps mitigate this. If you don't want to fill L1, try to keep L2 continuous.

[Abdella @MohAbdella]

@NERZ_z - Space out tracks to reduce crosstalk.
- Tracks on L1 shouldn't cross the splits on L2.
- Fill top with GND plane & stitching vias.
- Clear silkscreen interference with display outline.
- Move silkscreen of the sensor at the corner as it will be covered when sensor is assembled.

[Bernie Sanders @BernieSanders]

This is not a normal statement from a normal cabinet member of a major nation. 

This statement from Israeli National Security Minister Ben-Gvir is the statement of a war criminal. 

The racist, extremist Israeli government does not deserve one nickel of U.S. support.

[Abdella @MohAbdella]

@raheelys Trying to figure it out

[Abdella @MohAbdella]

@raheelys @princeofprawns0 Yes skidl isn't promising... Following your journey I hope you figure it out soon

[Abdella @MohAbdella]

@AlperenAkkuncu Tried it for you. Looks like opus one shot it correctly

[Abdella @MohAbdella]

@offline3alatool @mostafaezztaman كان ممكن يعمل اكتر؟ طيب لو بالاحتمالات فكان برضو ممكن يعمل اقل.. كان ممكن مايلاقيش شغل.. كان ممكن يشتغل و مديره او زمايله او المواصلات تنكد عليه.. كان ممكن ... كان ممكن؟ اللي عمله دا الرزق المكتوب له و هو انجاز ممتاز تبارك الله

[Graeme @gkisokay]

The Local LLM Cheat Sheet for your 32GB RAM device

I was asked to put together a practical lineup of local models that fit comfortably on a 32GB machine.

At this tier, you start getting access to real flagship-class local models, plus a growing number of custom quants. But for most people, these are the core models worth knowing first.

Flagship Models

Qwen3.5 27B / GGUF / Q6_K_M
The best overall 32GB flagship. General chat, writing, research, and agent workflows. Great if you want one model that can handle almost everything well.

Qwen3.6-35B-A3B / GGUF / UD-Q4_K_M
Best MoE flagship. Stronger for coding, reasoning, and tool use than most smaller generalists.

Gemma 4 31B / GGUF / Q6_K_M
Dense premium model. Writing, analysis, reasoning, and high-end local chat. Heavier than the MoE options, but excellent when quality matters more than speed.

Models for Fast Flagship Use

Gemma 4 26B A4B / GGUF / Q6_K_M
Great balance of speed and quality for general assistant work, coding, agent tasks, and research. This is one of the best 32GB picks if you want something that feels high-end without dragging.

DeepSeek-R1 Distill Qwen 32B / GGUF / Q4_K_M
Offline reasoning engine. Best for math, logic, deliberate analysis, and step-by-step problem solving.

Mistral Small 24B / GGUF / Q6_K_M
Tool-calling specialist. Strong for assistants, chat workflows, local business tasks, and function calling. Available for 24GB machines.

Models for Companion Use

Qwen3.5 9B / GGUF / Q6_K_M
Best sidekick. Fast drafts, search loops, cheap retries, and secondary agent work. Even on a 32GB machine, you still want a smaller model around for support tasks.

Llama 3.1 8B / GGUF / Q6_K_M
Long-context companion. RAG, doc ingestion, codebase chat, and long prompts. The output quality is not the sharpest anymore, but it is still useful when needing simple tasks fast.

From what my community tells me, the best single models are Qwen3.5 27B or Gemma 4 31B.

For two models, the strongest general pairing is Qwen3.5 27B + Qwen3.5 9B.

If you are more code-heavy, Qwen3.6-35B-A3B + Llama 3.1 8B.

Let me know what models you are running on 32GB, and which ones have actually been worth the RAM.

[Abdella @MohAbdella]

@AlperenAkkuncu Google's image recognition seems to be ahead of Claude and that's why it correctly extracted the parts and values from the pdf/image. The reason Claude is hallucinating in your case is likely the incorrect extraction of data so it got confused but not the reasoning itself.

[Abdella @MohAbdella]

@AlperenAkkuncu Yes, give it the netlist and BOM (even if partially ready). If you're also providing a pdf or image for the schematics, ask it to cross-check them with the netlist

[Abdella @MohAbdella]

@xzcendence @lauriewired FPGA with embedded Arm core?