HackerTrans
トップ新着トレンドコメント過去質問紹介求人

canopi

no profile record

投稿

MCPMark: A LLM Benchmark based on real-world use cases (in Notion, Playwright..)

mcpmark.ai
1 ポイント·投稿者 canopi·7 か月前·0 コメント

コメント

canopi
·6 か月前·議論
I still don't understand how memes work. The approach itself in this case is trivial, I think it's easy to argue that there are 10s of similar algorithms for using AI coding agents that would result in very similar results. However somehow the weird name + an initial cabal made this into a meme.
canopi
·6 か月前·議論
I did not know that digg still exists. I still remember the great move to reddit after the UI changes.

It's a bummer it requires a signup to use.
canopi
·7 か月前·議論
I would. I use pgadmin and feel it's very outdated and over-complicated for simple use cases.
canopi
·7 か月前·議論
Congrats on the launch. The problem is definitely there. I wonder how are you planning to differentiate yourself from Cursor and the like. You mention you are complementary, but Cursor provide similar features to add external doc context for instance to a prompt. I understand you do better in your benchmark, but with the amount of funding they may be able to replicate and improve over it (unless you have a secret thing).
canopi
·7 か月前·議論
Funny thing I remembered. OpenAI was one of the design partners for wandb. Funny they ended up using then acquiring a competitor.
canopi
·7 か月前·議論
Many thoughts:

- That wasn't in my bingo list. I guess it's acquisition time and all these AI companies need to spend the money somewhere.

- Congrats to the team. I was a big fan of their devrel work.

- Now both wandb and neptune AI have been acquired. I guess there is no independent MLOps solution left.

- Are they shutting down the product? Is this an acquihire? Does it mean OpenAI has not offices in Europe?
canopi
·7 か月前·議論
That's really cool. I guess this is an alternative to ngrok (which I like but hate due to having to sign in).
canopi
·7 か月前·議論
If that's not a proof of the 10/90 rule in machine learning. The last 10% of accuracy are harder than the first 90 (and that goes recursively).

We almost solved OCR 20 years ago. Then we spent 20 years on the last percentage. We see the same in self-driving cars.
canopi
·7 か月前·議論
This.

There is so much spam from random people about meaningless issues in our docs. AI has made the problem worse. Determining the meaningful from the meaningless is a full time job.
canopi
·7 か月前·議論
The first thing that comes to my mind is SOC2 HIPAA and the whole security theater.

I am one of the engineers that had to suffer through countless screenshots and forms to get these because they show that you are compliant and safe. While the real impactful things are ignored
canopi
·7 か月前·議論
Quite lucky for the projects that are trending :D