Hacker News: kennethwolters

Litelm: LiteLLM Without the Bloat

kennethwolters — Fri, 11 Sep 2026 18:10:20 +0000

Article URL: https://github.com/kennethwolters/litelm

Comments URL: https://news.ycombinator.com/item?id=49662767

Points: 174

# Comments: 60

New comment by kennethwolters in "Litelm: Litellm Without the Bloat"

kennethwolters — Tue, 24 Mar 2026 20:08:11 +0000

First commit was 2 weeks ago. A clean-room rewrite of only the most essential APIs of litellm.

Was annoyed that DSPy was about 250 MB big. Mostly due to litellm being bloated and numpy. In order to create dspy-lite [1], I had to create a lighter litellm and make numpy optional (only essential for embeddings).

Made the lighter litellm litelm by selecting the parts of litellm's test suite that mattered and build it back to match litellms API.

Litelm: Litellm Without the Bloat

kennethwolters — Tue, 24 Mar 2026 20:08:11 +0000

Article URL: https://github.com/kennethwolters/litelm

Comments URL: https://news.ycombinator.com/item?id=47508371

Points: 1

# Comments: 1

New comment by kennethwolters in "AI 2.0"

kennethwolters — Thu, 08 Jan 2026 19:31:30 +0000

Wrote a piece arguing that AI 1.0 is feudal because GPU scarcity concentrated power in a few cloud providers, but the coming GPU glut will invert this dynamic and make open weights + commodity compute the foundation of AI 2.0. The thesis is that sovereignty requires ownership of your stack (weights, compute, data) rather than renting intelligence from centralized APIs.

AI 2.0

kennethwolters — Thu, 08 Jan 2026 19:31:30 +0000

Article URL: https://kennethwolters.com/posts/ai2/

Comments URL: https://news.ycombinator.com/item?id=46545353

Points: 3

# Comments: 1

New comment by kennethwolters in "Ask HN: What Are You Working On? (December 2025)"

kennethwolters — Sun, 14 Dec 2025 21:54:55 +0000

LLM-driven narrative game. Main technical issue is how go do compaction. I’ve devised a memory hierarchy that compacts the story to a constant amount of tokens per layer. Arc -> Scene -> Moment -> Line. Not sure if that’s the right dimensions to decompose into. Also tinkering how to get the right amount of “divergence” for story progression option generation. A lot of unanswered questions…

New comment by kennethwolters in "Ephemeral infrastructure: Why short-lived is a good thing"

kennethwolters — Fri, 05 Dec 2025 16:31:20 +0000

for me it feels like: Everything is stateful by default/convenience. Building robust systems is in part about confining statefulness to as few parts as possible. To contain statefulness. It’s to buy you some time and capacity. Yet the toughest problems often arise in the stateful parts of the system as well as quasi-stateless parts which sometimes develop hidden statefulness (think of syncing webclient and server state). So being good at handling stateful systems is valuable. Maybe one should even embrace statefulness. However, the AWS Solution Architect will tell you otherwise.

New comment by kennethwolters in "AI Under the Hood Part I: Understanding the Machine"

kennethwolters — Fri, 03 Oct 2025 11:39:32 +0000

LLMs are pretty slow. Started a blog post series sharing what I've learned in search of faster LLMs.

AI Under the Hood Part I: Understanding the Machine

kennethwolters — Fri, 03 Oct 2025 11:39:32 +0000

Article URL: https://kennethwolters.com/posts/ai-under-hood-0/

Comments URL: https://news.ycombinator.com/item?id=45461722

Points: 4

# Comments: 2

New comment by kennethwolters in "No AGI in Sight: What This Means for LLMs"

kennethwolters — Tue, 12 Aug 2025 10:40:11 +0000

Wrote an essay on what LLMs reaching their technological limits could mean to "AI Engineering".

No AGI in Sight: What This Means for LLMs

kennethwolters — Tue, 12 Aug 2025 10:40:11 +0000

Article URL: https://kennethwolters.com/posts/no-agi/

Comments URL: https://news.ycombinator.com/item?id=44874578

Points: 4

# Comments: 1

New comment by kennethwolters in "Buy European Made. Support European Values"

kennethwolters — Thu, 06 Mar 2025 11:45:59 +0000

not an attempt at rebutting or so, just adding my thoughts:

It doesn't matter if the product is great because of those values if there is a lack product (bc politics and bureaucracy make it infeasible to produce). Makes you think if those values might contribute to the bad politics and bureaucracy. And if so, how.

New comment by kennethwolters in "Llama-OCR: Document to Markdown"

kennethwolters — Sat, 16 Nov 2024 13:14:00 +0000

Why don't you think that calling the app "Llama-OCR" is good enough?

New comment by kennethwolters in "Claude for Desktop"

kennethwolters — Thu, 31 Oct 2024 16:03:17 +0000

can someone tell me what the difference is to the webclient? Seems like a webview to me. No local caching/store.

New comment by kennethwolters in ""Tinyboxes finally have a buy it now button""

kennethwolters — Tue, 27 Aug 2024 12:36:16 +0000

Yeah that's a problem. I agree.

New comment by kennethwolters in ""Tinyboxes finally have a buy it now button""

kennethwolters — Tue, 27 Aug 2024 12:29:21 +0000

You are responsible for curating your "For You" tab by following/liking-posts-of/replying-to-posts-of/muting/blocking/unfollowing the right accounts.

Once I accepted that my "For You" tab got much better.

New comment by kennethwolters in "Flux: Open-source text-to-image model with 12B parameters"

kennethwolters — Thu, 01 Aug 2024 16:54:21 +0000

It is very good at "non-human subjects in photos with shallow focus".

Really curious to see what other low-hanging fruits people are finding.

New comment by kennethwolters in "Txtai: Open-source vector search and RAG for minimalists"

kennethwolters — Mon, 22 Jul 2024 16:17:31 +0000

It is very impressive :)

New comment by kennethwolters in "Txtai: Open-source vector search and RAG for minimalists"

kennethwolters — Mon, 22 Jul 2024 14:44:26 +0000

txtai is not (just) a vector store, it's a full-fledged RAG system. Apples and oranges if you ask me.

New comment by kennethwolters in "Txtai: Open-source vector search and RAG for minimalists"

kennethwolters — Mon, 22 Jul 2024 14:38:14 +0000

Don't have anything to add to the others. Just sharing a way of thinking for deciding between RAG and fine-tuning:

(A) RAG is for changing content

(B) fine-tuning is for changing behaviour

It's a bit simplistic but I found it helpful so far.