Burak Yigit Kaya
@byk.im
Curious mind. Open source, behavioral psychology, automation
And we are going Bun-first. If we can find a way to shring the Bun binary sizes, I don't think we'll have any reason to go with npx at this point!
February 2, 2026 at 9:59 AM
And we are going Bun-first. If we can find a way to shring the Bun binary sizes, I don't think we'll have any reason to go with npx at this point!
Then we needed SQLite and semver operations inside the CLI and to my pleasant surprise Bun has *BOTH* built-in. So now we have this `node-polyfills` file for npx/node support: github.com/getsentry/cl...
github.com
February 2, 2026 at 9:59 AM
Then we needed SQLite and semver operations inside the CLI and to my pleasant surprise Bun has *BOTH* built-in. So now we have this `node-polyfills` file for npx/node support: github.com/getsentry/cl...
This is a v0 launch and we'll be iterating quickly. Please send us your feedback so we can make it work great for you!
@maditya.sh, Miguel (bete), and I are eager to hear all from you!
@maditya.sh, Miguel (bete), and I are eager to hear all from you!
January 29, 2026 at 8:23 PM
This is a v0 launch and we'll be iterating quickly. Please send us your feedback so we can make it work great for you!
@maditya.sh, Miguel (bete), and I are eager to hear all from you!
@maditya.sh, Miguel (bete), and I are eager to hear all from you!
We took our inspiration from the awesome `gh` CLI from @github.com folks and gave it a Sentry spin with some vibe coding 😉
January 29, 2026 at 8:23 PM
We took our inspiration from the awesome `gh` CLI from @github.com folks and gave it a Sentry spin with some vibe coding 😉
Possibly. How I see this: the way LLMs work can be done deterministic as it is all mathematical functions. Eventhough it relies on probabilites, you can still deterministically control what is the cut off and use no random system. So when this becomes a business need, it is possible to get done.
December 4, 2025 at 6:26 PM
Possibly. How I see this: the way LLMs work can be done deterministic as it is all mathematical functions. Eventhough it relies on probabilites, you can still deterministically control what is the cut off and use no random system. So when this becomes a business need, it is possible to get done.
While looking for "hybrid" tips (silicon outer with foam inside) I came by this post: www.reddit.com/r/headphones...
And boy it blew my mind!
And boy it blew my mind!
From the headphones community on Reddit: I decided to make my own Hybrid IEM foam/silicon tips, works amazingly well. Made a guide for anyone who wants to try it. Best of both worlds!
Explore this post and more from the headphones community
www.reddit.com
December 4, 2025 at 6:21 PM
While looking for "hybrid" tips (silicon outer with foam inside) I came by this post: www.reddit.com/r/headphones...
And boy it blew my mind!
And boy it blew my mind!
Great article thanks!
>We reject this defeatism. With a little bit of work, we can understand the root causes of our nondeterminism and even solve them!
From a user's perspective, it's just temperature. Rest of the non-deterministic behavior comes from optimizations.
>We reject this defeatism. With a little bit of work, we can understand the root causes of our nondeterminism and even solve them!
From a user's perspective, it's just temperature. Rest of the non-deterministic behavior comes from optimizations.
December 2, 2025 at 2:36 PM
Great article thanks!
>We reject this defeatism. With a little bit of work, we can understand the root causes of our nondeterminism and even solve them!
From a user's perspective, it's just temperature. Rest of the non-deterministic behavior comes from optimizations.
>We reject this defeatism. With a little bit of work, we can understand the root causes of our nondeterminism and even solve them!
From a user's perspective, it's just temperature. Rest of the non-deterministic behavior comes from optimizations.
Determinism is also solved, especially with the temperature.
December 1, 2025 at 12:01 PM
Determinism is also solved, especially with the temperature.
This is exactly what's needed to make the next productivity jump. Not an agent helping you plan trips or whatever (still useful just not very profitable). The shift will come when we start doing work and thinking about it in a completely new way. Especially for white collar jobs. No more BS jobs.
December 1, 2025 at 11:00 AM
This is exactly what's needed to make the next productivity jump. Not an agent helping you plan trips or whatever (still useful just not very profitable). The shift will come when we start doing work and thinking about it in a completely new way. Especially for white collar jobs. No more BS jobs.
Both agree and disagree. Agree with the observation, disagree that this is a problem. Anthropic models are not just good at coding, they are good at problem solving. I've seen the models try things and then realize they are stuck and unblock themselves numerous times.
December 1, 2025 at 11:00 AM
Both agree and disagree. Agree with the observation, disagree that this is a problem. Anthropic models are not just good at coding, they are good at problem solving. I've seen the models try things and then realize they are stuck and unblock themselves numerous times.
Agree with Google for the end user side. That said I find Gemini consistently worse compared to ChatGPT. Amazon would win through platform hosting. I still don't see who's filling the enterprise model gap except for Anthropic though.
December 1, 2025 at 10:52 AM
Agree with Google for the end user side. That said I find Gemini consistently worse compared to ChatGPT. Amazon would win through platform hosting. I still don't see who's filling the enterprise model gap except for Anthropic though.
They don't need to as the real money will be from businesses. And TBH Claude is killing it for me compared to ChatGPT. Sonnet 4.5 is an incredibly good model.
December 1, 2025 at 10:34 AM
They don't need to as the real money will be from businesses. And TBH Claude is killing it for me compared to ChatGPT. Sonnet 4.5 is an incredibly good model.