Matt Silverlock
@elithrar.bsky.social
“Who do we say the rules are for?” “Other people.” • VP of Product: Databases & Storage @cloudflare • https://github.com/elithrar • https://twitter.com/elithrar
I wanted a way to benchmark a bunch of LLMs from the CLI.
Wrapped @simonwillison.net’s llm library + had it ingest a config file specifying the models you want. It'll run them async (if the plugin supports it) and write the results out to JSON so you can review/eval/process.
Wrapped @simonwillison.net’s llm library + had it ingest a config file specifying the models you want. It'll run them async (if the plugin supports it) and write the results out to JSON so you can review/eval/process.
February 3, 2025 at 9:56 PM
I wanted a way to benchmark a bunch of LLMs from the CLI.
Wrapped @simonwillison.net’s llm library + had it ingest a config file specifying the models you want. It'll run them async (if the plugin supports it) and write the results out to JSON so you can review/eval/process.
Wrapped @simonwillison.net’s llm library + had it ingest a config file specifying the models you want. It'll run them async (if the plugin supports it) and write the results out to JSON so you can review/eval/process.
When one of the PMs on the team *builds a macOS app* that uses Workers AI to auto-transcribe notes from any system process, summarize them, and do it all privately within your own Cloudflare account 😎
January 10, 2025 at 10:38 PM
When one of the PMs on the team *builds a macOS app* that uses Workers AI to auto-transcribe notes from any system process, summarize them, and do it all privately within your own Cloudflare account 😎
More reading in 2025, starting with these:
January 1, 2025 at 2:35 PM
More reading in 2025, starting with these:
Case in point: I like AI for prototyping, but… are we headed towards a monoculture because that’s all our tools support?
Do other libraries stand a chance? Do we get to try out new ideas?
I’m hoping the next big advance in LLMs enables more real-time knowledge (that isn’t just function calling).
Do other libraries stand a chance? Do we get to try out new ideas?
I’m hoping the next big advance in LLMs enables more real-time knowledge (that isn’t just function calling).
December 29, 2024 at 7:43 PM
Case in point: I like AI for prototyping, but… are we headed towards a monoculture because that’s all our tools support?
Do other libraries stand a chance? Do we get to try out new ideas?
I’m hoping the next big advance in LLMs enables more real-time knowledge (that isn’t just function calling).
Do other libraries stand a chance? Do we get to try out new ideas?
I’m hoping the next big advance in LLMs enables more real-time knowledge (that isn’t just function calling).
Swapped from iTerm 2 -> Ghostty. Super minimal config. Liking it so far: it really does just work + feel native.
December 27, 2024 at 4:05 PM
Swapped from iTerm 2 -> Ghostty. Super minimal config. Liking it so far: it really does just work + feel native.
When the docs engineering team picks the only correct date format:
December 23, 2024 at 11:56 AM
When the docs engineering team picks the only correct date format:
Half crispy (11 mins), half chewy (12.5 mins) 👨🍳
December 21, 2024 at 4:56 PM
Half crispy (11 mins), half chewy (12.5 mins) 👨🍳
When Hyperdrive launched (Sept 2023) / where we are today (Dec 2024):
December 11, 2024 at 11:41 PM
When Hyperdrive launched (Sept 2023) / where we are today (Dec 2024):
The Cloudflare Hyperdrive team just casually shipping a ~90% performance improvement via a changelog update:
December 11, 2024 at 11:40 PM
The Cloudflare Hyperdrive team just casually shipping a ~90% performance improvement via a changelog update:
Perfect Saturday listen: podcasts.apple.com/us/podcast/2... (Spider-Man 2 was the better movie)
December 7, 2024 at 4:47 PM
Perfect Saturday listen: podcasts.apple.com/us/podcast/2... (Spider-Man 2 was the better movie)
Just added the (brand new!) support for Llama 3.3 70B in WorkersAI to llm-cloudflare
$ llm install llm-cloudflare
$ llm -m "@cf/meta/llama-3.3-70b-instruct-fp8-fast" <PROMPT>
$ llm install llm-cloudflare
$ llm -m "@cf/meta/llama-3.3-70b-instruct-fp8-fast" <PROMPT>
December 7, 2024 at 1:37 AM
Just added the (brand new!) support for Llama 3.3 70B in WorkersAI to llm-cloudflare
$ llm install llm-cloudflare
$ llm -m "@cf/meta/llama-3.3-70b-instruct-fp8-fast" <PROMPT>
$ llm install llm-cloudflare
$ llm -m "@cf/meta/llama-3.3-70b-instruct-fp8-fast" <PROMPT>
Per-hostname pricing is here: www.cloudflare.com/plans/ - scroll down (admittedly not easy to find):
December 3, 2024 at 3:02 PM
Per-hostname pricing is here: www.cloudflare.com/plans/ - scroll down (admittedly not easy to find):
93 games played and 18% of folks got it first try. Not bad, but just wait until I publish the next Workers trivia game 😎
November 30, 2024 at 11:39 PM
93 games played and 18% of folks got it first try. Not bad, but just wait until I publish the next Workers trivia game 😎
Eight years in this country and finally going to a Waffle House:
November 30, 2024 at 5:35 PM
Eight years in this country and finally going to a Waffle House:
Had to make sure I remembered the answers:
November 29, 2024 at 12:44 PM
Had to make sure I remembered the answers:
Currently have a tab open for a thread from 2009 with photos for a car-related project on my list for next weekend!
November 27, 2024 at 1:21 PM
Currently have a tab open for a thread from 2009 with photos for a car-related project on my list for next weekend!
I’ll never tire of this skyline 🗽
November 27, 2024 at 1:14 PM
I’ll never tire of this skyline 🗽
The @cloudflare.social Workers free plan just got an upgrade 😎
November 26, 2024 at 4:04 PM
The @cloudflare.social Workers free plan just got an upgrade 😎
I think this is the most astute framing of Hyrum's Law (www.hyrumslaw.com) that I've seen to date:
November 21, 2024 at 4:03 PM
I think this is the most astute framing of Hyrum's Law (www.hyrumslaw.com) that I've seen to date:
One of my favorite JS libraries is p-map (github.com/sindresorhus...) — it allows you to concurrently map over Promises, rate limit them & handle failure.
I find it especially useful when writing @cloudflare.social Workflows: each invocation of the mapper creates a unique step on-the-fly:
I find it especially useful when writing @cloudflare.social Workflows: each invocation of the mapper creates a unique step on-the-fly:
November 21, 2024 at 2:22 PM
One of my favorite JS libraries is p-map (github.com/sindresorhus...) — it allows you to concurrently map over Promises, rate limit them & handle failure.
I find it especially useful when writing @cloudflare.social Workflows: each invocation of the mapper creates a unique step on-the-fly:
I find it especially useful when writing @cloudflare.social Workflows: each invocation of the mapper creates a unique step on-the-fly:
Example: Queues has the concept of a broker (control plane) and then independently scaling storage shards (data plane) using Durable Objects: blog.cloudflare.com/how-we-built...
November 21, 2024 at 9:41 AM
Example: Queues has the concept of a broker (control plane) and then independently scaling storage shards (data plane) using Durable Objects: blog.cloudflare.com/how-we-built...
My holiday season as a Cloudflare employee:
November 20, 2024 at 11:46 AM
My holiday season as a Cloudflare employee:
Post a perfect album from the 90s that isn't Nirvana, Pearl Jam, Soundgarden, or Alice In Chains.
November 19, 2024 at 9:53 PM
Post a perfect album from the 90s that isn't Nirvana, Pearl Jam, Soundgarden, or Alice In Chains.
Still using @zed.dev as my daily driver. I am (of course) curious how they plan to drive revenue long-term and foster a sustainable extension ecosystem vs. VS Code.
Also keen to see how the AI assistant integration evolves: I think that's the winning part but the space moves *fast*.
Also keen to see how the AI assistant integration evolves: I think that's the winning part but the space moves *fast*.
November 18, 2024 at 1:54 PM
Still using @zed.dev as my daily driver. I am (of course) curious how they plan to drive revenue long-term and foster a sustainable extension ecosystem vs. VS Code.
Also keen to see how the AI assistant integration evolves: I think that's the winning part but the space moves *fast*.
Also keen to see how the AI assistant integration evolves: I think that's the winning part but the space moves *fast*.