Ashok Gelal
@ashokgelal.bsky.social
Dev + Founder crafting the ultimate LLM playground @msty.app.
Building innovative products, fueled by coffee and inspired by raising two future legends. Code is life, and life is beautiful!
Building innovative products, fueled by coffee and inspired by raising two future legends. Code is life, and life is beautiful!
Gemini is good (and mostly free). It has a larger context window so that's helpful for big docs.
"Claude requires you to type "continue" every couple minutes."
Some apps like @msty.app as Continue built-in. That way you don't have to type and the subsequent responses are "stitched" together.
"Claude requires you to type "continue" every couple minutes."
Some apps like @msty.app as Continue built-in. That way you don't have to type and the subsequent responses are "stitched" together.
April 9, 2025 at 1:37 PM
Gemini is good (and mostly free). It has a larger context window so that's helpful for big docs.
"Claude requires you to type "continue" every couple minutes."
Some apps like @msty.app as Continue built-in. That way you don't have to type and the subsequent responses are "stitched" together.
"Claude requires you to type "continue" every couple minutes."
Some apps like @msty.app as Continue built-in. That way you don't have to type and the subsequent responses are "stitched" together.
I hear ya! What model(s) are you using?
April 8, 2025 at 12:50 PM
I hear ya! What model(s) are you using?
Ya, some Reddit posts saying the same thing. It’d be a sad day if I moved on though. I advocate for it everywhere I go, and always make my work pay for it (even though I already have a personal license), I have fought with my colleagues about it and converted at least a few. Filed a ticket, will see
February 9, 2025 at 12:49 PM
Ya, some Reddit posts saying the same thing. It’d be a sad day if I moved on though. I advocate for it everywhere I go, and always make my work pay for it (even though I already have a personal license), I have fought with my colleagues about it and converted at least a few. Filed a ticket, will see
It has gone slower over the time. I have enabled Use types from server as well (as recommended on YT and Reddit). Memory is not an issue (dedicated 64GB!! just for WS). Will DM you. This is painful so please help 😭
February 8, 2025 at 9:26 PM
It has gone slower over the time. I have enabled Use types from server as well (as recommended on YT and Reddit). Memory is not an issue (dedicated 64GB!! just for WS). Will DM you. This is painful so please help 😭
Any chance you are on Discord? If not, my handle at gmail (or X DM if you have an account there)
January 30, 2025 at 2:37 PM
Any chance you are on Discord? If not, my handle at gmail (or X DM if you have an account there)
When I read the news, the first thing that came to mind was you and knew that you'd be stoked hearing this. And I was also wondering if you still had that watch or not. Glad to hear that it is still there somewhere with you :)
January 28, 2025 at 8:04 PM
When I read the news, the first thing that came to mind was you and knew that you'd be stoked hearing this. And I was also wondering if you still had that watch or not. Glad to hear that it is still there somewhere with you :)
Happy to jump on a call and chat more if you want
January 26, 2025 at 6:41 AM
Happy to jump on a call and chat more if you want
Almost all of our users use multiple models so I think letting users rank them would be better.
Some model providers do provide both models and chat interface so I think models and "model provider's chat interface" should be two different things. Not sure if that is helpful or more confusing.
Some model providers do provide both models and chat interface so I think models and "model provider's chat interface" should be two different things. Not sure if that is helpful or more confusing.
January 26, 2025 at 6:41 AM
Almost all of our users use multiple models so I think letting users rank them would be better.
Some model providers do provide both models and chat interface so I think models and "model provider's chat interface" should be two different things. Not sure if that is helpful or more confusing.
Some model providers do provide both models and chat interface so I think models and "model provider's chat interface" should be two different things. Not sure if that is helpful or more confusing.
And I feel like there needs to be a dedicated section for “best LLM Models” that would include Claude, chatGPT, O1, DeepSeek etc. Happy to talk more and contribute.
January 25, 2025 at 11:47 PM
And I feel like there needs to be a dedicated section for “best LLM Models” that would include Claude, chatGPT, O1, DeepSeek etc. Happy to talk more and contribute.
Chat apps maybe? I wrote-in to Chatbots. Also, calling OpenAI, Anthropic, etc. chat bots felt weird. As someone could use OpenAI, Claude etc. with an app like Msty. FWIW, we call them Online AI Model providers. Then you can fit in other providers such as AWS Bedrock which don’t have chat interface
January 25, 2025 at 11:47 PM
Chat apps maybe? I wrote-in to Chatbots. Also, calling OpenAI, Anthropic, etc. chat bots felt weird. As someone could use OpenAI, Claude etc. with an app like Msty. FWIW, we call them Online AI Model providers. Then you can fit in other providers such as AWS Bedrock which don’t have chat interface
Reposted by Ashok Gelal
Even better with Roo-Cline but Msty and tools like it (LM Studio etc.) are game changes in barrier to entry ✨️
January 19, 2025 at 11:56 AM
Even better with Roo-Cline but Msty and tools like it (LM Studio etc.) are game changes in barrier to entry ✨️