Open Source
440 subscribers
157 photos
2 videos
222 files
189 links
Download Telegram
๐Ÿš€ Paid Internship Opportunity โ€“ Backend Developer (Node.js / PostGIS)

AmbaLay Maps is building local mapping & location intelligence infrastructure for Ethiopia and Africa.

Weโ€™re looking for a Backend Developer Intern who wants to learn by building real systems, not just tutorials.

What youโ€™ll work on
- Node.js backend APIs
- PostgreSQL / PostGIS
- Location & map-based data

Who should apply
- CS / Software / IT students or recent graduates
- Node.js knowledge
- Curious, disciplined, and willing to learn
(GIS experience is a plus, not required)

Apply here or send your CV / GitHub to:
jobs@ambalaymaps.com

Join: link
๐Ÿ‘5๐Ÿ”ฅ1
Open Source pinned ยซhttps://t.me/boost/et_open_sourceยป
Professional CV.pptx
108.5 KB
It was a fruitful training session!
Hope, next time you apply for remote Jobs with a professional CV๐Ÿ˜‰

Source :maroset
Join:link
๐Ÿ‘1
Join:link
๐Ÿ‘1
Join: link
Join : link
โค2๐Ÿ‘1
Join: link
โค1๐Ÿ‘1๐Ÿ‘1
Open Source
Photo
DR. Milan

๐—ฆ๐—ฒ๐—ป๐—ถ๐—ผ๐—ฟ ๐—ฑ๐—ฒ๐˜ƒ๐˜€ ๐—ฎ๐—ฐ๐—ฐ๐—ฒ๐—ฝ๐˜๐—ฒ๐—ฑ ๐—”๐—œ ๐—ฐ๐—ผ๐—ฑ๐—ฒ ๐˜€๐˜‚๐—ด๐—ด๐—ฒ๐˜€๐˜๐—ถ๐—ผ๐—ป๐˜€ ๐—บ๐—ผ๐—ฟ๐—ฒ ๐—ผ๐—ณ๐˜๐—ฒ๐—ป ๐˜๐—ต๐—ฎ๐—ป ๐—ท๐˜‚๐—ป๐—ถ๐—ผ๐—ฟ๐˜€

A University of Chicago study tracked 44,160 Cursor users over eight months.

The findings flip conventional assumptions about AI coding tools.

๐—ง๐—ต๐—ฒ ๐—ฒ๐˜…๐—ฝ๐—ฒ๐—ฟ๐—ถ๐—ฒ๐—ป๐—ฐ๐—ฒ ๐—ฝ๐—ฎ๐—ฟ๐—ฎ๐—ฑ๐—ผ๐˜…

Senior engineers accept agent-generated code 6% more often for every standard deviation increase in experience. Junior engineers use agents more frequently but take less of what they generate.

This reverses the pattern with autocomplete tools, where juniors accept more suggestions.

๐—ช๐—ต๐˜† ๐˜€๐—ฒ๐—ป๐—ถ๐—ผ๐—ฟ๐˜€ ๐˜„๐—ถ๐—ป ๐˜„๐—ถ๐˜๐—ต ๐—ฎ๐—ด๐—ฒ๐—ป๐˜๐˜€

The study identified three factors:

1. Senior engineers write tighter specifications with less ambiguity. They break work into agent-compatible units.

2. They have stronger heuristics for evaluating correctness, making code review faster and more accurate.

3. Experienced developers are also more likely to instruct agents to plan before implementing. This planning step improves alignment between what the agent produces and what the developer actually wants.

๐—ง๐—ต๐—ฒ ๐—ฝ๐—ฟ๐—ผ๐—ฑ๐˜‚๐—ฐ๐˜๐—ถ๐˜ƒ๐—ถ๐˜๐˜† ๐—ป๐˜‚๐—บ๐—ฏ๐—ฒ๐—ฟ๐˜€

After Cursor's agent feature became the default, software output increased 39%. No significant changes in short-term quality metrics, such as revert rates, were observed over the six-month window.

Cross-sectionally, frequent agent users merge more code and work on larger-scoped changes.

๐—ช๐—ต๐—ฎ๐˜ ๐˜๐—ต๐—ถ๐˜€ ๐—บ๐—ฒ๐—ฎ๐—ป๐˜€

Agents don't replace engineering skill. They amplify it. The developers who get the most value are those who can clearly specify what they want, decompose problems effectively, and evaluate output quickly.

The paper suggests three skills matter more in an agent-assisted workflow:

๐Ÿ”น Abstraction (understanding what agents can do)
๐Ÿ”น Clarity (specifying intent precisely)
๐Ÿ”น Evaluation (assessing output against requirements)

If you're struggling with AI coding tools, the issue might not be the tool. It might be that effective agent use requires skills that traditionally mattered less when you were typing every line yourself.



Join : link
โค2