nishantmodak

I built a small MCP server that lets any AI agent (Claude Desktop, Cursor, Windsurf) block and wait for a human response before taking irreversible actions.

  Two tools:
  - ask_human(question) – agent blocks until you reply
  - request_approval(action) – agent blocks until you click Approve or Deny

  Notifications via Slack (with interactive buttons), or native macOS
  dialogs. Works with any MCP-compatible client over stdio or SSE.

  The use case that motivated it: I kept coming back to find my agent had done
  something I didn't want while I was away. Now it just asks.

  GitHub: https://github.com/nishantmodak/call-a-human-mcp

Problem has never been the storage. Its running those queries to return in milliseconds - if its for a dashboard, an alert of your new AI agent trying to make sense of it.

I wrote this after seeing the same failure modes across multiple teams and metrics stacks.

The intent here isn’t to argue that high cardinality is bad but to explain the mechanical reasons unbounded cardinality shows up later as cost surprises, slow queries, and trust erosion.

Curious if others have seen similar query-time or indexing issues they have noticed.

Wasn't it pascal who said that all of our problems come from our inability to be bored.

I have a different take

Engineers have to pre-define and send all telemetry data they might need – since it’s so difficult to make changes after the fact – regardless of the percentage chance of the actual need.

YES. Let them send all the data. The best place to solve for it is at Ingestion.

There's typically 5 different stages to this process.

Instrumentation -> Ingestion -> Storage -> Query (Dashboard) -> Query (Alerting)

Instrumentation is the wrong place to solve this.

Ingestion - Build pipelines that allow to process this data and provide for tools like streaming aggregation, cardinality controls that allow to 'process it' or act on anomalous patterns. This atleast makes working on observability data 'dynamic' instead of having to go change instrumentation always. Storage - Provide blaze (2hours), hot(1 month), cold(13 months) of tiered data storage with indipendent read paths.

This, in my opinion has solved for the bulk of cost & re-work challenges associated with telemetry data.

I believe, Observability is the Big Data of today, without the Big Data tools! (Disclosure: I work at Last9.io and we have taken a similar approach to solve for these challenges)

https://last9.io/data-tiering/

Link https://blog.last9.io/sre-tooling-clever-hans-fallacy/

I can see this in my inbox too!

The rice farmers I met in Bihar, for instance, are now growing a new variety of flood-tolerant rice—nicknamed “scuba” rice—that can survive two weeks underwater.

Found this extremely interesting. Probably because, I was surprised by the thought that, this could be done

Agree! With http://tah.io/get we have tried to make use of BLE not just as beacons but also for combining the power of your physical sensors with 'smartphone' sensors to have much more interesting use cases!

(Disclaimer: Work at TAH) I totally agree to what you are saying. We have had a similar experience working with our devices and SDK.

http://tah.io/get

You should also include dates alongside the breach. Just so that one knows, if they have taken corrective action after that

Interesting to see bitly stats on the link https://bitly.com/retweetdeals+

$10 million . 10k startups.

$1000 in Google Adwords + Google Apps credits.

This is what they mean!

“If we had had the same progress in the aircraft industry, you and I could be flying between London and New York in 500,000-seat planes and the fare would be about 25 cents,” he once said.

What kind of progress/innovation does he refer to in this?

Having played around with reveal.js (http://lab.hakim.se/reveal-js/) just last week. This is even more interesting!

Thanks for the pointer to original discussion. Some interesting points on that one.

Isn't it just easy to Opt-Out from Online Behavioral Advertising..

http://www.aboutads.info/choices/

I think someone's just playing around using his name

Look at the profile

http://disqus.com/guest/8f1dc8d0613be097f161d62bf5488744/

https://www.tin-nsdl.com/pan-verify/pan-verification-overvie...

Yes. It can be. Since the NSDL service is available to banks.

And since this program works in conjunction with ICICI. it's very possible

It's peer pressure. http://en.m.wikipedia.org/wiki/Peer_pressure

When I see this - I really appreciate how good 'Watson' is with speech to text!

Totally agree. I wondered the same. May be - since he is a _famous_ hacker - an update about his location might help those in the area with communities / talks etc?

http://news.ycombinator.com/item?id=1207721 < Why writing software is not like engineering.

I wanted this thread to a single place for people to list what they can offer.

In China, Baidu for sure!

I back-up most of my important files as an email attachment.

is it just me who does this way?

and the headline reading 'Europe Betting on Self-Regulation to control the internet'

What was awesome is - when i searched for something and it did not yield many results. As the last link, it suggested try this search on google. that's just gr8

this sounds entirely like 'connecting the dots' from Steve Jobs famous speech.

Looks like, Australia is the place to be! :)