CODE

→ View original post on X — @avikumart_

Building Your Technical Identity Across Developer Platforms

By

@avikumart_

–

12 January 2026 19h09

I have had trouble establishing my technical identity across various platforms, including GitHub, Kaggle, & others If you’re a developer/researcher, your "real" resume is buried in GitHub commits, arXiv, and X. Recruiters/collaborators usually only see a fraction of what you do

→ View original post on X — @miketamir

Code Execution with MCP: Building Efficient AI Agents

By

@miketamir

–

12 January 2026 17h32

Code execution with MCP: building more efficient AI agents Anthropic https://
buff.ly/1QcODUL
#AI #MachineLearning #DeepLearning #LLMs #DataScience

Optimizing AI Generation: Zero Failure Rate and Reduced Compute Costs

By

–

12 January 2026 16h39

5/5 What that unlocked: – Failure rate: 30% → 0
– Repo downloads: 8,000+ → 500
– Provisioning: per run → once per campaign
– And the “don’t burn compute” part: split Generation (patch) from Evaluation (tests). If evaluation fails, we retry tests, not re-generate tokens.

Optimizing SWE-bench Performance Through Shared Resource Provisioning

By

–

12 January 2026 16h39

4/5 The fix was a mindset shift: stop provisioning per run. Provision per SWE-bench instance, then reuse it. Shared per instance: repo at the right commit, MCP server, dependencies. Unique per run: execution path, files changed, commands, patch. We keep runs isolated with

SWE-bench Evaluation Challenges at Kubernetes Scale

By

–

12 January 2026 16h39

3/5 Trying to run SWE-bench eval as-is on k8s at large scale wasn't trivial: – Fresh pods have no cache. This means that everyone re-downloads the world (hello HF 429s.
– “docker run inside k8s” works on paper, then dies from contention, privileges, and overhead. It worked, but

Agentic Evaluation: An Infrastructure Challenge, Not a Model Problem

By

–

12 January 2026 16h39

1/5 We ran SWE-bench 200,000+ times to get statistical confidence in agentic evals. The main lesson wasn’t about prompts or models.
It was: agentic evaluation is an infrastructure problem.

→ View original post on X — @sumanth_077

LLaMA-Factory: Fine-Tune 100+ LLMs Without Coding

By

@sumanth_077

–

12 January 2026 14h54

Fine-Tune 100+ LLMs without writing a single line of code! LLaMA-Factory lets you train and fine-tune open-source LLMs and VLMs without writing any code. Here's why it's a game changer for fine-tuning: • Fine-tune 100+ LLMs/VLMs with built-in templates (LLaMA, Gemma, Qwen,

→ View original post on X — @dfintelligence

Linus joins the vibe coding sect, code is worthless

By

@dfintelligence

–

12 January 2026 13h59

Linus has joined the vibe coding sect. Code is definitely worthless. I love it.

→ View original post on X — @dfintelligence

Tests and corrections on separate branches, less than 5% errors

By

@dfintelligence

–

12 January 2026 10h45

Yes, totally. With the right tests running afterwards and you ask it to correct until the tests pass, well, developing on a separate branch each time is fine. Sure, some things slip through the cracks, I won't lie. But it's less than 5%, so.