Shivam Dwivedi

Shivam Dwivedi

Oct 27

What the VKORE NPU release lets you check

#machinelearning #ai

2

4 min read

Shivam Dwivedi

Oct 27

DecisionFormer has two confidence fields. Check which one you are thresholding.

#machinelearning #python #ai

5

4 min read

Shivam Dwivedi

Oct 27

A Hugging Face download should leave a receipt

#huggingface #python

1

3 min read

Shivam Dwivedi

Oct 26

PaperPilot runs locally. Your agent still has a data boundary.

#ai #privacy #opensource

3

4 min read

Shivam Dwivedi

Oct 26

Sixteen JAX devices on a CPU, and one gradient bug worth finding

#python #machinelearning #tutorial

5

4 min read

Shivam Dwivedi

Oct 26

The first LLM router metric I check is the missing-row count

#ai #machinelearning #python

4 min read

Shivam Dwivedi

Oct 25

LitPivot makes a quiet research bug visible: the metric can outlive the idea

#ai #machinelearning #research

5

4 min read

Shivam Dwivedi

Oct 25

Claude can write the proof. You still have to choose the theorem.

#python #ai #security

5

4 min read

Shivam Dwivedi

Oct 25

The OpenSpec update I want first is the one that writes nothing

#ai #tutorial #opensource

5

4 min read

Shivam Dwivedi

Oct 24

Mundo puts a simulator inside the world model. Here’s what to inspect.

#simulation #python #ai

5

5 min read

Shivam Dwivedi

Oct 24

Kimi K3.1 Pro: 104B active parameters still carry a 2.8T backpack

#python #ai #machinelearning

5

5 min read

Shivam Dwivedi

Oct 24

MCP Tasks: the retry bug hiding behind a successful job

#mcp #python #tutorial

5

2

5 min read

Shivam Dwivedi

Oct 23

ThinkingBox’s 20/20 column is useful. It still isn’t a reliability guarantee.

#machinelearning #ai

4

3 min read

Shivam Dwivedi

Oct 23

OpenShell’s default policy is only the start of the access story

#ai #security #architecture

3 min read

Shivam Dwivedi

Oct 23

Bun has a formatter now. Start with the command that writes nothing.

#bunjs #javascript #tutorial

3

3 min read

Shivam Dwivedi

Oct 22

Before copying Google's Rust rules, read the scope labels

#rust

5

3 min read

Shivam Dwivedi

Oct 22

TypeScript 7.1 can check an import attribute and still miss a missing file

#typescript #javascript

4

3 min read

Shivam Dwivedi

Oct 22

Claude Duet makes “should I interrupt you?” a product feature

#news #ai

5

3 min read

Hacktoberfest Open-Source AI Challenge Week 3: Slow and Steady Submission 🦥

Shivam Dwivedi

Oct 21

Sloth Inbox: let a tiny open model sort the notes you keep avoiding

#devchallenge #hf26challenge #ai #javascript

5

5 min read

Shivam Dwivedi

Oct 21

DeepInfer is open. Here is where the engine stops.

#ai #opensource #machinelearning #architecture

3 min read

Shivam Dwivedi

Oct 21

Try tsrs without replacing your TypeScript build

#typescript #rust #tutorial #webdev

5

3 min read

Shivam Dwivedi

Oct 20

GPT vs DeepSeek vs Gemini Argon: one ball, three animations

#showdev #ai #javascript #testing

5

5 min read

Shivam Dwivedi

Oct 20

10x faster generation, 1.4x faster workflow: a Python lab

#python #performance #asyncio #ai

6 min read

Shivam Dwivedi

Oct 20

DeepSeek Flash 1017 skips Jinja. Read its prompts with Python

#ai #python #opensource #tutorial

4

6 min read

Shivam Dwivedi

Oct 20

Free MiniMax, zero API keys: build a read-only SQL report app

#ai #python #sqlite #tutorial

5 min read

Shivam Dwivedi

Oct 20

Mercury 3 is fast. Your agent still has to wait for its tools.

#ai #machinelearning #programming #performance

5

3 min read

Shivam Dwivedi

Oct 20

DeepSeek Flash 1017: check the alias, the harness and the checkpoint

#news #ai #opensource #programming

3 min read

Shivam Dwivedi

Oct 20

MiniMax M3.1 Flash is Space Bunny. Here’s what you can actually use.

#news #ai #programming #agents

4

3 min read

Shivam Dwivedi

Oct 19

Python returned zero. The coin model still assigned a positive probability.

#python #statistics #math #scipy

5

1

7 min read

Shivam Dwivedi

Oct 18

The cookie jar behind A/B testing: Fisher’s test to churn

#statistics #datascience #python #scipy

5

12 min read

Shivam Dwivedi

Oct 18

Probabilities can sum to one and still describe the wrong experiment

#python #scipy #statistics #math

5

7 min read

Shivam Dwivedi

Oct 17

OpenJev vs OpenAI Decisions: fewer missed defects, more false alarms

#machinelearning

5

10 min read

Shivam Dwivedi

Oct 17

OpenJev on text and images: a tie, then a scoring surprise

#ai #machinelearning #python #opensource

5

6 min read

DEV Community

Badges

Writing Debut

Three Year Club

Two Year Club

One Year Club

What the VKORE NPU release lets you check

DecisionFormer has two confidence fields. Check which one you are thresholding.

A Hugging Face download should leave a receipt

PaperPilot runs locally. Your agent still has a data boundary.

Sixteen JAX devices on a CPU, and one gradient bug worth finding

The first LLM router metric I check is the missing-row count

LitPivot makes a quiet research bug visible: the metric can outlive the idea

Claude can write the proof. You still have to choose the theorem.

The OpenSpec update I want first is the one that writes nothing

Mundo puts a simulator inside the world model. Here’s what to inspect.

Kimi K3.1 Pro: 104B active parameters still carry a 2.8T backpack

MCP Tasks: the retry bug hiding behind a successful job

ThinkingBox’s 20/20 column is useful. It still isn’t a reliability guarantee.

OpenShell’s default policy is only the start of the access story

Bun has a formatter now. Start with the command that writes nothing.

Before copying Google's Rust rules, read the scope labels

TypeScript 7.1 can check an import attribute and still miss a missing file

Claude Duet makes “should I interrupt you?” a product feature

Sloth Inbox: let a tiny open model sort the notes you keep avoiding

DeepInfer is open. Here is where the engine stops.

Try tsrs without replacing your TypeScript build

GPT vs DeepSeek vs Gemini Argon: one ball, three animations

10x faster generation, 1.4x faster workflow: a Python lab

DeepSeek Flash 1017 skips Jinja. Read its prompts with Python

Free MiniMax, zero API keys: build a read-only SQL report app

Mercury 3 is fast. Your agent still has to wait for its tools.

DeepSeek Flash 1017: check the alias, the harness and the checkpoint

MiniMax M3.1 Flash is Space Bunny. Here’s what you can actually use.

Python returned zero. The coin model still assigned a positive probability.

The cookie jar behind A/B testing: Fisher’s test to churn

Probabilities can sum to one and still describe the wrong experiment

OpenJev vs OpenAI Decisions: fewer missed defects, more false alarms

OpenJev on text and images: a tie, then a scoring surprise