Hacker Newsnew | past | comments | ask | show | jobs | submit | runeb's commentslogin

Very cool! I was just today thinking I should have an AI help me design a new piece of furniture I've been thinking about, and send it off to a CNC shop.

The MCP server seems to have a bug. The tool list flashes for a second in Claude Desktop, but then immediately states "This connector has no tools available." and the error "Couldn’t reload tools from the server." when I try to refresh it.


Thanks for reporting the issue, it should be fixed now. Make sure to fully remove and then re-connect the MCP, and let me know what you think! If you like you can describe what you're building here and I'll make a demo of how I would prompt it.


In the last few months I have been building an AI assistant that is quite possibly saving my life.

It's an iOS app that uses LLMs to estimate and track nutrients and calories in all the food I eat. It is plugged in to Apple Health so it sees all my activity, sleep, body metrics, etc. I have set some general health goals in the app and it uses this to rate (poor, neutral, good) and advise me on the food I log. A calendar view shows how each day has gone in a traffic light coloring system. I can also log contextual notes to help it make sense of my day, and it has a memory system so each session has continuity.

The assistant can save extracted or inferred nutritional data in a library, like when I give it photos of nutritional labels for foods I eat frequently or I log the same thing repeatedly, for easy reuse later. Also useful for home made meals and recipes.

I've connected it to my workout tracker app Hevy so it can keep tabs on my exercise and also adjust my programs based on me chatting with it in the app about how a workout went. This is really useful for swapping out exercises I struggle with, for example, or to keep a well-balanced program over time.

Recently I added an Apple Watch companion app with a watch face complication so I can keep an easy eye on things like up-to-date calorie deficit.

All built in Swift with Opus on the $20 plan. I realize I can probably use existing AI nutrition tracking apps, and probably also connect Claude to Apple Health, but the joy of building and getting this to be completely tailored to me has been so fun and extremely useful. I've shared it with a few friends, but since I am paying for all the API tokens currently I have not yet tried to get more users onboarded.


Wow, I’ve had it for 15 years and until now I’ve never thought it may be connected to also being «treated» for glue ear as a kid. Do you have more info on this and how you came to that conclusion?


(I see you use another quoting style, so I'll try to keep my language to the point) I only heard about the term tinnitus as a 'teen or perhaps later.

ENT: Ear, Nose and Throat - a medical specialism.

My earliest memories are from around 18 months. I used to have very bad earache whenever I had a cold or influenza. My mother told me that I used to burst into tears, unprovoked or without any of the usual baby related reasons and eventually around age five or six I was diagnosed with "glue ear".

The way it was explained to me was: my eustacian tubes were too narrow. Speaking to children: "Your ears, nose and throat are all linked together and the tubes that link your ears to the rest of you are too small and get blocked easily".

I had surgery to insert "grommets" into my eustacian tubes. This was done twice. Grommets are tiny plastic, hollow tubes. From memory: about 3mm long and 3mm outer diameter but that is from a long time ago.

Now here is where it gets complicated! My dad was in the British army (so was my mum but that's another story). This means we moved house every two years or so. I can fix dates quite well.

The first operation was performed in Rinteln, West Germany at a British Forces hospital. We were stationed in Paderborn so it would be around 1976. I had a second set inserted in Wythenshawe Hospital in Manchester (UK) in 1977 or perhaps early 1978.

I know that the second set of grommets were removed by a doctor (I think an ENT specialist) but I can't remember if the first set were removed in Manchester and replaced by the second pair at the same time.

I recall that the doctor used a black, plastic, bell shaped thing that fitted over my ear and wiggled some sort of hooked, metal probe within my ear to pull out the grommets. It was quite painful.

Even after the operations, I still got earache whenever I had a cold up until around age 20-25. That age range is a bit hard to pin down. I'll also note that when we were stationed in Cyprus in 1986-7ish, a lot of swimming and diving helped clear the tubes!

One of the nasty side effects of glue ear is that you have trouble with pressure changes. Airliners and swimming are the bane of your life.

There are some notes. Hope it helps.


The bane of your life, yet helped clear the tubes. Interesting


So it was you! /s


A particularly pronounced version of this can often be seen by letting 2 agents review and code in a loop. One agent will find some problems with the code, the other agent will address the review by adding more code.

A good human developer might see that the better way to address the review is to backtrack and pick a different approach. The ai agents seem more prone to getting stuck down bad branches of the decision tree.


Code reviews should be done by someone other than the author though, so the only thing that changes with ai generated code in that respect is the amount of it


That doesn’t conflict with my statement.

Before: One person writes the code (and likely understands it thoroughly), another person reviews the code to spot obvious mistakes or shortcomings. Now: AI writes the code, a person reviews it to spot obvious mistakes or shortcomings.

In the before case, you have a person who has a deeper understanding of the code and in the AI case, you don’t, instead you have even more code to review.

When a competent programmer is writing the code, the human written code tends to be higher quality too. So it’s not just about review quantity but the quality of code being reviewed. Some people claim the AI writes great code, but that just hasn’t been my experience yet (at least with the models I’ve tried, including Opus). They still make ridiculously bad decisions regularly.


>When a competent programmer is writing the code, the human written code tends to be higher quality too

This is a great idea, but on average is deeply untrue. Far and away most programmers today write significantly worse code than LLMs. Also LLMs are fantastic at generating high level summaries and comments in code


> Far and away most programmers today write significantly worse code than LLMs

Your experience with LLMs do not match my own. Not to say that I haven’t experienced terrible human written code where I’ve wondered what the author could possibly have been thinking, but overall, I still find LLM written code to be on the poor side.

Like, the code itself is ok, but the wider picture reasoning and abstractions are bad. It also makes really dumb decisions far too often. Or doggedly shoehorns its first idea in no matter how badly it fits.


They specifically mention "H.264, H.265, and av1 codecs, along with many others" here https://red.anthropic.com/2026/mythos-preview/


Checkout tailscale


And what about a website I want to make public? I'm just concerned about my IP being visible, like for my personal website or my searxng instance


Personally I'd just proxy it through a vm running on hertzer, linode, rackspace, etc


Tailscale Funnel, but might need a paid account


Tailscale's control plane uses Cloudflare.


Thanks, I did not know this. My Tailscale was unaffected by the outage.


I read their point as why have passwords at all when the security is you having access to your email account.


> The backup roll thrusters in use are now at risk due to residue buildup in their fuel lines

Such a human experience this probe is having


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: