Just saw this post elsewhere and it's so damn good
November 12, 2025 at 5:12 AM
Just saw this post elsewhere and it's so damn good
I repeated the same test and it got first shot 👌
September 28, 2025 at 3:31 PM
I repeated the same test and it got first shot 👌
#mogzi v0.14.1 released (alpha)
- Token usage & context usage tracking greatly improved
- Sessions naming now includes fruits & ice cream names
🐞: parallel tool calling is being wonky
🐞: shell tool needs handle interactive cmds better
#ai-engineeing #dotnet #csharp #microsoft #programming
- Token usage & context usage tracking greatly improved
- Sessions naming now includes fruits & ice cream names
🐞: parallel tool calling is being wonky
🐞: shell tool needs handle interactive cmds better
#ai-engineeing #dotnet #csharp #microsoft #programming
July 15, 2025 at 2:18 PM
#mogzi v0.14.1 released (alpha)
- Token usage & context usage tracking greatly improved
- Sessions naming now includes fruits & ice cream names
🐞: parallel tool calling is being wonky
🐞: shell tool needs handle interactive cmds better
#ai-engineeing #dotnet #csharp #microsoft #programming
- Token usage & context usage tracking greatly improved
- Sessions naming now includes fruits & ice cream names
🐞: parallel tool calling is being wonky
🐞: shell tool needs handle interactive cmds better
#ai-engineeing #dotnet #csharp #microsoft #programming
Introducing #maxbot - (private beta)
- cross-platform, open source, terminal, **multi-provider**...
- Simply built with the amazing Microsoft.Extensions.AI, Spectre.Console, and Native AOT 🎁
- Supports STDIN pipe-ling, file system security, etc
#dotnet #charp #ai #terminal #linux #osx #microsoft
- cross-platform, open source, terminal, **multi-provider**...
- Simply built with the amazing Microsoft.Extensions.AI, Spectre.Console, and Native AOT 🎁
- Supports STDIN pipe-ling, file system security, etc
#dotnet #charp #ai #terminal #linux #osx #microsoft
July 2, 2025 at 6:31 AM
I've been cooking 🔥...coming soon.
- cross-platform, open source, terminal, **multi-provider**...
- Simply built with the amazing Microsoft.Extensions.AI, Spectre.Console, and Native AOT 🔥
- Supports STDIN pipeling, file system security, etc
#dotnet #charp #ai #terminal #linux #osx #microsoft
- cross-platform, open source, terminal, **multi-provider**...
- Simply built with the amazing Microsoft.Extensions.AI, Spectre.Console, and Native AOT 🔥
- Supports STDIN pipeling, file system security, etc
#dotnet #charp #ai #terminal #linux #osx #microsoft
July 1, 2025 at 1:14 PM
My two 2003 Mac Pro Keyboards finally went kaput. After an extensive search, I purchased the BOYI K104 Pro Mechanical Keyboard. Bro, I am loving the look, the feel, and the sound.
www.amazon.com/BOYI-Mechani...
www.amazon.com/BOYI-Mechani...
April 6, 2025 at 2:46 AM
My two 2003 Mac Pro Keyboards finally went kaput. After an extensive search, I purchased the BOYI K104 Pro Mechanical Keyboard. Bro, I am loving the look, the feel, and the sound.
www.amazon.com/BOYI-Mechani...
www.amazon.com/BOYI-Mechani...
I think they hired Dr. Doofenshmirtz for all product branding 😂
April 3, 2025 at 9:36 PM
I think they hired Dr. Doofenshmirtz for all product branding 😂
Case in Point about correct Doctors
March 23, 2025 at 10:52 PM
Case in Point about correct Doctors
Ah. I read the rest of your thread. Yeah I did the ollama local thing and it was a none starter. I'm happy to pay providers to use the same OSS LLMs but with insanely faster inference for a couple of bucks.
Also the smaller LLMs didn't fair so we'll in this paper
Also the smaller LLMs didn't fair so we'll in this paper
March 23, 2025 at 2:44 AM
Ah. I read the rest of your thread. Yeah I did the ollama local thing and it was a none starter. I'm happy to pay providers to use the same OSS LLMs but with insanely faster inference for a couple of bucks.
Also the smaller LLMs didn't fair so we'll in this paper
Also the smaller LLMs didn't fair so we'll in this paper
Results of LLM performance of medical questions. Y-axis is hallucination resistant (higher is better). X-asis is how similar the LLM answer is to the correct answer.
March 23, 2025 at 12:51 AM
Results of LLM performance of medical questions. Y-axis is hallucination resistant (higher is better). X-asis is how similar the LLM answer is to the correct answer.
I don't think you read this study correctly. The headline is coming from an "opinion survey" of only 70 reapondants. But when they performed their own tests, Deepseek R1 score a ~90%🤘 for correct answers and 90%🤘 for hallucinatio. Resistance 👈
"Doctors don't know WTF they're talking about."
"Doctors don't know WTF they're talking about."
March 23, 2025 at 12:36 AM
I don't think you read this study correctly. The headline is coming from an "opinion survey" of only 70 reapondants. But when they performed their own tests, Deepseek R1 score a ~90%🤘 for correct answers and 90%🤘 for hallucinatio. Resistance 👈
"Doctors don't know WTF they're talking about."
"Doctors don't know WTF they're talking about."
Conversational #AI Chat with #claude via requesty.ai and #dotnet ~50 lines of #csharp 😎
This example assumes an OpenAI-Compatible provider to access the LLM.
Kudos to the #microsoft team building Microsoft.Extensions.AI🤖
This example assumes an OpenAI-Compatible provider to access the LLM.
Kudos to the #microsoft team building Microsoft.Extensions.AI🤖
March 6, 2025 at 1:39 AM
Conversational #AI Chat with #claude via requesty.ai and #dotnet ~50 lines of #csharp 😎
This example assumes an OpenAI-Compatible provider to access the LLM.
Kudos to the #microsoft team building Microsoft.Extensions.AI🤖
This example assumes an OpenAI-Compatible provider to access the LLM.
Kudos to the #microsoft team building Microsoft.Extensions.AI🤖
A lot of vets voted for this draft dodging loser. Vets, snap out of it, swallow your pride and join in the march to end MAGA.
March 5, 2025 at 6:39 PM
A lot of vets voted for this draft dodging loser. Vets, snap out of it, swallow your pride and join in the march to end MAGA.
Results from an experiment utilizing a beta feature built by requesty.ai specifically for users of #cline and #roocode.
The beta features intercept and manipulate the chat stream on its way to the LLMs. In this case, trimming portions of the system project coming from #ai tools
#aiengineering
The beta features intercept and manipulate the chat stream on its way to the LLMs. In this case, trimming portions of the system project coming from #ai tools
#aiengineering
March 5, 2025 at 4:46 AM
Results from an experiment utilizing a beta feature built by requesty.ai specifically for users of #cline and #roocode.
The beta features intercept and manipulate the chat stream on its way to the LLMs. In this case, trimming portions of the system project coming from #ai tools
#aiengineering
The beta features intercept and manipulate the chat stream on its way to the LLMs. In this case, trimming portions of the system project coming from #ai tools
#aiengineering
March 5, 2025 at 12:51 AM
Why I will always go with Blazor 🎉
March 4, 2025 at 7:34 AM
Why I will always go with Blazor 🎉
...and I found your post using my handy new bluesky
February 18, 2025 at 5:11 AM
...and I found your post using my handy new bluesky