je42

> All participants were then reassured that though 20 minutes might be a rather short time to write an essay, they were encouraged to do their best.

Given that the task has been under time pressure, I am not sure this study helps gauging the impact of LLMs in other contexts.

When my goal is to produce the result for a specific short term task - I maximize tool usage.

When my goal is to improve my personal skills - I use the LLM tooling differently optimizing for long(er) term learning.

I find that LLM extremely good in training such language skills by using following process:

a) write a draft yourself.

b) ask the LLM to correct your draft and make it better.

c) newer LLMs will explicitly mention the things they corrected (otherwise ask for being explict about the changes)

d) walk through each of the changes and apply the ones you feel that make the text better

This helped me improving my writing skills drastically (in multiple languages) compared to the times where I didn't have access to LLMs.

Looking at the co routine example. Does the function type implicitly change when using co.sleep in the function body?

One alternatives I like is https://bakpakin.github.io/tiny-ecs/doc/

There are alternatives that are not on the EU list for controlled substances. Like for example:

https://en.wikipedia.org/wiki/Lisdexamfetamine

More data can help inform semi closed loop systems to deal with this variability.

For example, AAPS has since version 3.2 dynamic IFS. ( https://androidaps.readthedocs.io/en/latest/DailyLifeWithAap... )

For me this works quite well

I wonder why Godot wasn't picked. Did I miss the points in the article?

Roblox introduced LuaU to get static typing

When using dynamic languages, either minimize code dependencies / function calls and complexity or ensure high test coverage.

Echarts can correctly animate data thats streaming via a moving window from right to left. Some other libs have issues in this case.

Good point. It is my experience with both of my teenage kids, diagnosed with adhd.

In general, it needs to be broad since it is difficult to predict if you get enough attention or not.

So starting with a very short explanation helps increasing the odds of that it will be processed.

Explainations need to be 2-3 short sentences. Longer, and even teenagers with Ahdh may not follow or retain the explanation offered.

I find that most LUA code sits between functional, imperative and OOP.

What works well is to measure the same food multiple times in a container (every time same volume ).

after a while you can estimate carbs by visually inspecting the contents without scale.

Add to that automatic bolus by a semi closed loop system to correct for errors, you can achieve good results with minimal effort.

For closed loop better cgms will help.

Main points are (fsl2 based):

- latency, currently 10min.

- accuracy, fine in normal range, but when you have a low blood sugar suddenly the latency spikes a lot.

- values when being under the shower too high.

- start up time of 60min could be lower.

-open up the hardware for any app to read

Love 2d can be quickly extended with hot reloading. See https://github.com/rxi/lurker

The incentives to deceive are the same for content creators regardless the platform. News outlet or YouTube or tiktok.

Joe Rogan vs CNN is a good example - that you can apply all the same doubts you have for Tiktok also to conventional news outlets.

However, if you dont have carb info, the alternative is to judge yourself. Your own model may be better than gpts model, though. I would use GPTs output and at least look at it on a case by case basis

Not only fat plays a role with pizza, but also the amount of protein in it. When having pizza we usually add protein to the carbs. 50% immediately bolus. Other 50% spread over 3-4 hours, and let AAPS dose the insulin.

My son (13) and I are also using AAPS for roughly 3 years now. The app does not look styled, however after taking the initial hurdle to find our way around in the app, it is greatly appreciated. Integrates nicely with Xdrip ( anf others)

Sometimes it is even a form field on Linkedin job openings.

E2E are less maintainable because by definition. They are the type of tests that has the most dependencies on sub-systems and other systems - this means if a test fails you ll need to work more to figure out what went wrong, than in an integration test, which depends on less sub-systems.

Nats also has some basic routing options. Which of RabbitMQ's options would give it the edge when deciding between the two?

I switched from iOS to Android for AAPS ( i didn't want to carry two phones )

With AAPS - it is usually able to deal with my blood sugars at night automatically.

During holidays ( 0 stress day), AAPS can almost completely determine boluses just based on my profile and carb intake. ( i.e. I don't force a bolus ) However, during normal workdays, I usually need to bolus right before carb intake. AAPS also allows you specify bolus a percentage and then let it figure out the rest automatically.

There is an AAPS-like app for iOS but I am not sure how good it is. See https://loopkit.github.io/loopdocs/#what-is-the-loop-app

I have Type 1 Diabetes (T1D) and use a semi-closed loop system, which includes the Libre 2 sensor with XDrip and the Omnipod Dash, managed by AAPS.

AAPS (unlike the Tandem T , which I’m less familiar with) uses a more sophisticated algorithm to calculate the required insulin, considering several factors:

- Blood sugar levels received via XDrip from the Libre 2 sensor.

- A profile that includes:

- - The basal rate per hour for each time slot

- - The insulin-to-carb ratio

- - The target blood sugar range

- Automatically calculated insulin sensitivity.

- Insulin on board (remaining active insulin), calculated using a curve to estimate how much insulin is still active after injection.

- Carbs on board, with AAPS tracking active carbs over time.

16kb of the 64kb was reserved for the screen buffer if i remember correctly.

Nice post ! I use a another set of tools to manage my type 1 and the one of my son.

Monitoring glucose:

Sensor libre 2 and xdrip mobile app using the bluetooth option where xdrip + oop2 reads values in 1min intervals. ( no miaomiao needed ) Values are pushed to nightscout

I got 2 nightscout instances: one for my self one for my son deployed to heroku.

Alarms are pushed to my phone for my sons low and high bloodsugar using pushover.

Heroku alarms are configured if service goes down.

Database: mongo atlas highly available.

We both use omnipod dash with android aps app.

A) helps with carb to insulin calc and carb counting, insuline on board.

B) closed loop, that auto corrects high glucose and stops insulin when too low

C) different profiles for things like "recently done sports" vs "didnt do much excerise in the last days"

D) aps app allows to track extra insuline injected by pen

Lazyvim can help. https://lazyvim.github.io/

> Transform your Neovim into a full-fledged IDE

Unfortunately not true. It is not a single metric.

we have extra rules on top: No doping, separation of genders,...

However, tools like buienrader seem to have trouble in the recent months/years to accurately predict local weather.