quan

I’m working on a proxy MCP server that lazy loads tools to save tokens https://github.com/mquan/nimble. It includes a dashboard for connecting and configuring MCP servers.

Reminds me of this old clip https://m.youtube.com/watch?v=3Lyex2tSUyA

During the holiday season in December, I got my 9yr old son into programming by making games with LLM. We made a few browser games, and by the end of it he could write the prompt himself, copy the code to the right places, understand what a function does and add objects to the game by modifying an array.

What I learned is llm gives you the activation energy, you just type a few sentences to get the momentum going. To get the reward feedback cycle going you’ll want to add graphics as early as possible.

The most difficult part is to be by his side and ready to jump in whenever a missing coma breaks the entire game. You’ll also need to keep scope under control, I’d alway steer him away from doing any complicated animation. There’re plenty of opportunity to teach but be realistic that it’s not real programming

Here’s an example that I put online so he can share with friends: https://mquan.github.io/k.ai/coin-collector/

afaik, the langchain solution loads entire openAPI spec which consumes a lot of token and won't work for many large API. For efficient token usage, api2ai divides the task into two steps: api planning and params parsing. First step takes a summarization of all the endpoints. Once the endpoint is known, we parse params using the schema of the selected endpoint.

I open sourced this tool that takes OpenAPI spec and let you control API using natural language https://github.com/mquan/api2ai

Let me know if you have any questions or feature request

I created api2ai (https://github.com/mquan/api2ai) to solve this. You provide an OpenAPI spec and auth data to spin up an agent for your API. It’s still low level and requires dev work. But I’m working on building tools to bring it to end users

I created api2ai, I agree that we want to get consistent result, but that can be solved by looking up cached AI results. If the input is user driven then you’ll need AI to decipher the natural language.

It’s done in two steps: 1. Create an array of every endpoint’s summary text and ask AI to select one based on the user prompt 2. Use function calling with parameters of the selected operation in step 1

Same as your dad, five of my uncles fought in the war on the U.S. side, my grandfather and granduncle served in the South Vietnamese government in high positions. My family was persecuted after the war, many fled, the ones that stayed spent time in prison and fled after.

In college, I read a lot on the Vietnam war and learned about the Tet Offensive with the Hue massacre by the Vietcong, the Land Reform prior to that. But I don't see how the photos at My Lai is "pure propaganda." If you spend some time to learn about the tragic event you will find that there were a big cover up and for over 500 human lives lost, only one man was charged and subsequently pardoned.

The U.S. committed to a war when it didn't understand the people or history of Vietnam. Vietnam has a long history of fighting foreign invaders, and most people perceive the U.S. as such as they are just replacing the French when they withdrew after Dien Bien Phu in 1954. Till this day, most Vietnamese people consider China to be their biggest enemy even though the two countries share the same political ideology.

Yes absolutely, I'm only curious because his essays inspire me to think differently about work and attracted me to this community. Thanks for your answer

What happened to pg, is he still involved in YC day to day operation or the application process?

If you google "nine-dash line", which violates UNCLOS, you'll understand why the ASEAN countries are upset and rightly so. China has visibly flexed their muscle in the region in recent years with building military bases, adding land to underwater islands so they can claim sea territory around them, and attacking fishing boats in disputed region. Some of this happen to dispute with Japan as well. With the economy not going well, I'm afraid this may be the Communist Party's strategy to diffuse tensions at home in order to hold on to power.

In addition, I find the claim crediting the Communist government for adopting Latin character to be bogus. If this is true, then south Vietnam would be a lot more illiterate than the north and not the thriving commercial and research center it is today, simply because these government-sponsored programs would be most effective in the northern part where the central government is. In fact, the Communist government was driven out into hiding in rural area as soon as they declared independence. So they wouldn't have the time or mean to implement country-wide education reforms. As one data point, my grandma, who did not come from privileged background and was born in 1920 and like most of her peers spoke and wrote French and Vietnamese fluently, received all her education prior to 1945. I'm pretty sure the Latin writing adoption was started much earlier and it was more natural simply because it's closest to French, which was the official language at the time

Is there any inflationary effect to the $1.25/day figure that has been used since the 1980s? It's not clear from the article whether it's adjusted for inflation.

My experience with Robinhood has been completely negative. I recently worked on a loan and needed to show the Robinhood transfer to my bank. Since there's no web interface you can't get the current month printout. There's no phone number to call, and it took them 3 days to reply to the zendesk ticket only to say they can't do anything and I have to wait til end of the month. The actual statement is only available a week after beginning of the month. Never again. I closed out my account and deleted the app.

I'd love to see this used on Fox News to fact check everything they say in real time.

I found this screencast by Tom Dale very comprehensive and extremely easy to follow http://www.youtube.com/watch?v=Ga99hMi7wfY

Just append image_url parameter like this: http://www.lavishbootstrap.com/?image_url=http://farm2.stati...

thanks, some intelligent way to automatically assign colors would be nice. May be I'll look at colors brightness and hue. For now, you can manually reassign them if the generated theme doesn't work.

Bad idea posting to HN with only 1 dyno, I cached the default color scheme so homepage should load a lot faster now.

Wow, what a coincidence! I just created a very similar project http://www.lavishbootstrap.com/

How did you generate the different themes, and are they stored in different css/less files?

I made a gem to do exactly this with RMagick, you can check it out here: https://github.com/mquan/prizm

I recently started looking for a new apartment and found it exhausting to keep track of all the posts. So I put this together to better visualize and compare apartments. It doesn't have to be about apartments though, you can edit the input text and play around with it.

I built this during the weekend as an experiment for my gistpoint app and open sourced the js code as a Raphael plugin here: https://github.com/mquan/polkadot.

If Arrington had invented Groupon, he would have invented Groupon

may be I'll let people specify a link to their twitter profile

Great feedbacks! Actually, it already measures article popularity using twitter count api. The articles shown on the frontpage are sorted by twitter link count.

If you install the bookmarklet and click on any page it will show you the machine extracted summary. You can click edit to submit your own summary.

Great idea to incentivize users to submit summaries by linking to their flattr profile.

thanks for the kind words, it's probably a problem with the fulltext extraction. If the fulltext is extracted correctly, the summary extraction should be good. Btw, I use the ruby-readability gem for the fulltext extraction.

I understand your point, the way I see people use this is to determine before committing their time if the article content fits their interests and worth their time. It's not to bypass reading articles altogether, it's to filter out what they want to read.

I know a font police is going to catch this crime, I hope it's appropriate for the posit notes :)