Skip to content
Navigation menu
Search
Powered by Algolia
Search
Log in
Create account
DEV Community
Close
#
llm
Follow
Hide
Posts
Left menu
đź‘‹
Sign in
for the ability to sort posts by
relevant
,
latest
, or
top
.
Right menu
OpenEval: Why LLM Evaluation Needs a Standard Format
Adha AK
Adha AK
Adha AK
Follow
Jul 30
OpenEval: Why LLM Evaluation Needs a Standard Format
#
llm
#
evaluation
#
ai
#
testing
Comments
Add Comment
1 min read
Making Local AI Tool Calls More Reliable
Alain Chan
Alain Chan
Alain Chan
Follow
Aug 21
Making Local AI Tool Calls More Reliable
#
agents
#
ai
#
llm
#
programming
2
 reactions
Comments
6
 comments
1 min read
Show HN: Telechat – Self-hosted Claude on Telegram/WhatsApp/Slack, no cloud relay
Subhendu Das
Subhendu Das
Subhendu Das
Follow
Jul 31
Show HN: Telechat – Self-hosted Claude on Telegram/WhatsApp/Slack, no cloud relay
#
showdev
#
ai
#
llm
#
opensource
2
 reactions
Comments
Add Comment
2 min read
Your agent returned 200 OK. Was it actually right?
Andrew Van Dyke
Andrew Van Dyke
Andrew Van Dyke
Follow
Aug 3
Your agent returned 200 OK. Was it actually right?
#
ai
#
llm
#
agents
Comments
Add Comment
2 min read
`finish_reason=length` Returned Empty Content — and the Error Message Lied to Me
Milo Bergstrom
Milo Bergstrom
Milo Bergstrom
Follow
Jul 30
`finish_reason=length` Returned Empty Content — and the Error Message Lied to Me
#
llm
#
debugging
#
typescript
#
ai
1
 reaction
Comments
Add Comment
3 min read
Model + Harness = Agent: The Gap Isn’t Where You Think
OctoLab
OctoLab
OctoLab
Follow
Jul 29
Model + Harness = Agent: The Gap Isn’t Where You Think
#
ai
#
agents
#
llm
#
programming
Comments
Add Comment
4 min read
I Trust My AI Completely—Except When It Says “Done”
OctoLab
OctoLab
OctoLab
Follow
Jul 29
I Trust My AI Completely—Except When It Says “Done”
#
ai
#
agents
#
llm
#
programming
1
 reaction
Comments
1
 comment
4 min read
Kmemo 2.0 is out, and the two gaps I admitted to in the first post are closed
AS
AS
AS
Follow
Aug 1
Kmemo 2.0 is out, and the two gaps I admitted to in the first post are closed
#
kotlin
#
llm
#
caching
#
opensource
4
 reactions
Comments
Add Comment
5 min read
I built an AI observability platform with $0 – zero dependencies, zero ops, stateless
Mamman Mohammed
Mamman Mohammed
Mamman Mohammed
Follow
Jul 29
I built an AI observability platform with $0 – zero dependencies, zero ops, stateless
#
showdev
#
ai
#
llm
#
software
2
 reactions
Comments
1
 comment
2 min read
The 22 Failure Modes Are Not 22 Problems. They Are Five.
Robert Evans
Robert Evans
Robert Evans
Follow
Jul 29
The 22 Failure Modes Are Not 22 Problems. They Are Five.
#
agents
#
ai
#
llm
#
machinelearning
1
 reaction
Comments
1
 comment
4 min read
Context compaction is silently destroying your LLM agent's memory
Shuo Liu
Shuo Liu
Shuo Liu
Follow
Aug 11
Context compaction is silently destroying your LLM agent's memory
#
llm
#
python
#
ai
#
opensource
Comments
8
 comments
3 min read
Deploying a QAT Checkpoint Your Serving Stack Can't Load: Gemma 4 E2B in Pure JAX on One TPU
xbill
xbill
xbill
Follow
for
Google Developer Experts
Aug 19
Deploying a QAT Checkpoint Your Serving Stack Can't Load: Gemma 4 E2B in Pure JAX on One TPU
#
tpu
#
jax
#
llm
#
quantization
8
 reactions
Comments
Add Comment
10 min read
How to make your Next.js site appear in ChatGPT (and any LLM)
Gustavo Garcia
Gustavo Garcia
Gustavo Garcia
Follow
Aug 3
How to make your Next.js site appear in ChatGPT (and any LLM)
#
chatgpt
#
llm
#
nextjs
#
seo
Comments
Add Comment
9 min read
The Same GraphRAG Comparison Wins and Loses. It Depends Which Instrument Judged It.
Edward Izgorodin
Edward Izgorodin
Edward Izgorodin
Follow
Aug 29
The Same GraphRAG Comparison Wins and Loses. It Depends Which Instrument Judged It.
#
ai
#
rag
#
llm
#
benchmarks
10
 reactions
Comments
23
 comments
3 min read
MCP moves context. It doesn't create it.
Taran Singhania
Taran Singhania
Taran Singhania
Follow
for
Sentra
Aug 21
MCP moves context. It doesn't create it.
#
mcp
#
ai
#
llm
#
architecture
Comments
1
 comment
5 min read
đź‘‹
Sign in
for the ability to sort posts by
relevant
,
latest
, or
top
.
We're a place where coders share, stay up-to-date and grow their careers.
Log in
Create account