manishsharan

7 ms·

31.
by manishsharan 1y ago
https://github.com/microsoft/graphrag This is not agentic but pretty good results when I did a poc.
32.
by manishsharan 1y ago
Most developers can't do much work without an IDE and Chrome + Google. Would you say that their work has no value?
33.
by manishsharan 1y ago
This is exactly my experience. We wanted to modernize a java codebase by removing java JNDI global variables. This is a simple though tedious task. And we tried Claude Code and Gemini. Both of these results were hilarious.
34.
by manishsharan 1y ago
I would also love to see how the model performs on management type presentations. I am referring to Gartner and McKinsey and BCG etc. presentations.
35.
by manishsharan 1y ago
Not always. You could paste your entire repo into Gemini and it could map your forest and also identify the "trees". Assuming your codebase is smaller than Gemini context window. Sometimes it makes sense to upload a package,s code
36.
by manishsharan 1y ago
This price drop is nice but I wonder how long it will last. Their prices used to be very low,then they almost doubled, and now it dropped.
37.
by manishsharan 1y ago
>>After all, after nazi-salute, mecha-hitler, etc blew up, by just how much did the demand for Teslas fall? I had been saving up for a Tesla but now I am looking elsewhere. I think a lot of people are doing the same here in Canada. Yo
38.
by manishsharan 1y ago
Not the above poster. I would like to run agents with local multimodal LLMs to process huge numbers of sensitive files for an org( summarization, knowledge extraction, answer user questions ,etc). Any ideas?
39.
by manishsharan 1y ago
Who is paying for those tokens? I had a long conversation with "Devin" and I must have burnt up a large number of tokens. In any case, thank you "Devin"
40.
by manishsharan 1y ago
Gemini and chatgpt and github copilot subscriptions also provide similar functionality.
41.
by manishsharan 1y ago
The next cheapest and capable model is GLM 4.5 at $0.6 per million tokens in and $2.2 per million tokens out. Glad to see DeepSeek is still be the value king. But I am sti disappointed with the price increase.
42.
by manishsharan 1y ago
I am not interested in the results of the 95%. I want to know more about the 5% who got it right. What are their use cases ?
43.
by manishsharan 1y ago
Does anyone else miss o3? I swear I had an understanding of how to get deep analytical thinking out of o3. I am absolutely struggling to get the same results with GPT-5. The new model feels different and frustrating to use.
44.
by manishsharan 1y ago
My apologies This is a better repo to learn about Claude code internals https://github.com/ghuntley/claude-code-source-code-deobfusc...
45.
by manishsharan 1y ago
I have been following along the code in this repo. https://github.com/ghuntley/claude-code-source-code-deobfusc... The author has done a pretty good job of reverse engineering Claude Code and explaining the architectur
46.
by manishsharan 1y ago
Fixing Vibe Code!
47.
by manishsharan 1y ago
Is there a reason why you have not added DeepSeek and Qwen and Meta ? You should also aggregate prices from Vertex and AWS Bedrock .
48.
by manishsharan 1y ago
Are they going to require my DNA and blood samples to access this ? Screw their organization verification. I am taking my business to Claude or Deepseek.
49.
by manishsharan 1y ago
Will this make Google mad at me and cancel/freeze all my Google services ?
50.
by manishsharan 1y ago
In my experience, Claude Code is scary good. Gemini CLI is just plain dumb and not worth the time.
51.
by manishsharan 1y ago
This 100%. Frameworks like these, and I include Microsoft Semantic kernel and Google' Agent2Agent, just get in the way. Fact is that Agent development is still in early stages. We daily learn more about how to get the most out of LLMs.
52.
by manishsharan 1y ago
I provided 'grep' as a tool to LLM (deepseek) and it does a better job of finding usages. This is especially true if the code is obfuscated JavaScript.
53.
by manishsharan 1y ago
Is there a tokenizer someone can recommend for code ? I have tried CodeBert but maybe I am using it wrong as my results with it were pretty bad.
54.
by manishsharan 1y ago
I actually read the actual study on Lancet. unfortunately for you, your vile prejudice can not be justified by this study.
55.
by manishsharan 1y ago
I decided to build a Agent system from scratch It is sort of trivial to build it. Its just User + System Prompt + Assistant +Tools in a loop with some memory management.. The loop code can be as complex as I want it to be e.g. I could snap
56.
by manishsharan 1y ago
Imagine capturing a crime on camera. And then having the court toss this as inadmissible evidence. Fun times ahead !
57.
by manishsharan 1y ago
Why not merely upload all relevant documents into Gemini? Split the knowledge into smaller knowledge domains and have agents ( backed by Gemini) for each domain?
58.
by manishsharan 1y ago
This is a lazy argument. It only serves to dissuade the masses from doing critical thinking and analyzing party platforms. It also explains why blue collar Americans vote for tax breaks for billionaires and union busting legislation.
59.
by manishsharan 2y ago
Have you notice more verbose posts in your feed ? Llama is allowing everyone to sound more knowledgeable than they are. AI based content generation is like an instragram filter for intellect; everyone is pretending to be thoughtful.
60.
by manishsharan 2y ago
Joy Con Drift. Can they please solve for that ? Chat Button ? Camera ? are they intentionally driving away parents who buy Switch for their kids ?

More ›