melony

I expect my "AI" engineers to know both RAG and what KL-divergences are. If you only know the former, you are just a backend developer, not a ML engineer.

How does it compare with NLTK's chunking library? I have found that it works very well for sentence segmentation.

There's nothing stopping them from just adding the feature like npm.

Some optimization suggestions:

- cache the prompting somehow, unless you are doing dynamic stuff with the prompts, the language embeddings generated should be static (this depends on the architecture of the model that you are using, it's only possible with certain setups where the language processing is a separate part in the pipeline)

- consider fine-tuning an img to img model with your current outputs instead of using a language-coupled model. My intuition is that this is currently significantly over-engineered on the ML side.

- Play around with local hardware acceleration instead of sending everything to the cloud, you also probably don't need particularly high resolution for the images either.

Sigh it means in their neural network layer they have a random sampling step. Yay for Gaussians.

Does it work with WSL2?

The government has power. Craig the office drone asking for "misinformation" to be taken down at behest of his boss does not have the power "The government" does. Or am I now supposed to believe that if you anger Sheila at the DMV the IRS will audit you?

Craig's boss is the highest executive power in the country. Most of their work is handled by aides and subordinates like Craig, with authority delegated from their office. The boss only directly makes the important decisions. You are being disingenuous by claiming Craig is a innocent, powerless trigger-happy drone. And if you have lived in small towns, expect to receive worse service or discrimination from Sheila in future. The highest power in the country cannot be held to the same standard as a county clerk (and frankly speaking, most DMV employees need more training in customer service and have their compensation tied to overall productivity and performance, lots of fat to be trimmed there, both literal and metaphorical).

The issue is that many things ultimately still require rote learning. Most STEM subjects cannot avoid it unless you are genuinely talented and mathematically inclined. You can appreciate Newton's laws in classical mechanics as much as you want but you still need practice if you want to apply the kinematics and force equations in any reasonable amount of time during an exam. Same with Hamiltonians and Lagrangians. The further you go, the more intuition requires rote learning to build up. (I would note however that rote learning doesn't necessarily impart any mathematical maturity, but it does help with achieving conceptual understanding of the problem. Solve enough differentiation problems on pen and paper and gradient descent and backpropagation become obvious, but proving epsilon deltas won't necessarily be as helpful.)

Because so long Bloom's 2 sigma observation holds, pedagogy only matters at scale. In other words, it mostly affect the lower middle class and poor people. Good education systems can ensure that most people get educated. The elites and upper middle class are more than capable of giving their children an excellent education anywhere. Working class parents don't have time to babysit every single homework problem or to fix every gap and flaw in the school's syllabus by themselves.

Why not use Go? They have one of the best networking stacks among language standard libraries, excellent support for low latency soft real-time concurrent operations, and single binary deployment is trivial.

Are there antennas in car wheel wells? The main car battery can probably power it with a correctly calibrated antenna.

The Gitea team is full of crypto shillers.

https://blog.gitea.io/2022/10/a-message-from-lunny-on-gitea-...

Their founder claims to have invented Gogs (he was one of the early committers) when the original author is another engineer

https://en.m.wikipedia.org/wiki/Gitea

Gitea was created by Lunny Xiao, who was also a founder of the self-hosted Git service Gogs.

https://about.sourcegraph.com/blog/three-years-at-sourcegrap...

About the author Joe Chen is Software Engineer and maintainer of the open source project Gogs, a painless self-hosted Git service. You can chat with Joe on Twitter @jc_unknwon or our community Discord

The Java web framework (Apache Wicket) it uses is vintage, about as old as Rails.

https://en.m.wikipedia.org/wiki/Apache_Wicket

OneDev: https://github.com/theonedev/onedev

Gogs: https://github.com/gogs/gogs

CyanogenMod?

What do also and use do?

My main gripe with Dart is their package manager. Unlike NPM or Cargo, it is possible to get conflicting diamond dependencies. The lack of shared memory concurrency is also disappointing.

I have seen fedora hat samurais at Comic con

They probably pulled another Winget. Their product manager will tease an acquihire and bingo your business's stolen

https://news.ycombinator.com/item?id=23331287

Indeed, what is Spanish for Aux armes, ingénieurs en développement logiciel?

They are going after it because it directly competes with Microsoft through piracy. Mojang provides (unofficially) Minecraft source code class maps for decompiled code through the Mod Coder Pack which is maintained by one of their staff. This is the foundation (client+server) modding frameworks like Minecraft Forge and Fabric are built on. Decompilation tools already exist and the Minecraft source code is an open secret. Their line in the sand is piracy at scale.

https://docs.minecraftforge.net/en/1.19.2/gettingstarted/#mi...

Turkish is another dreadfully regular language, would make for a great programming language

There was a company called reconfigure.io with a great solution where channel based concurrency was automatically compiled to parallel FPGA programs but they no long seem to be around

This is no different from the Samsung camera hallucinating the picture of the moon. This is not real signal processing in the traditional tomographic sense, it's taking the input and generating a posterior image conditioned on the input from an existing model.

The integrated debugger was also excellent. Outside of Lisp, few languages have decent support for Edit and Continue.

https://blog.codinghorror.com/edit-and-continue/

I think Java and .Net are pretty close. It is utterly shameful that languages like Python and JavaScript consider themselves to be developer friendly when their debuggers still require full reloads whenever the local state gets modified in place. Common Lisp had Edit and Continue (equivalent functionality in their debugger) before the dotcom crisis.

Vercel and Microsoft product managers, if you are reading this, please implement this feature in Next.js, VS Code, and Typescript.

The first thing that pops into your head when reading "private vehicle's location need not be streamed to government" is: why would it?

Wait till you hear about "airplanes".

Go into industry not academia. Academia won’t pay well and they do not appreciated good quality software engineering. For most labs, ML is considered a tool, not a primary driver of innovation. Look towards places like Isomorphic labs and similar companies that are cash rich and has a strong engineering heritage.

Any real innovation on the mechatronics side? The computer vision and controls are now increasingly dominated by deep learning.

Alon's a good reference but makes for very dry reading.

Be very very good at differential equations. Systems/synthetic biology primarily rely on dynamics to model behavior and side effects. AlphaFold is actually pretty different from the biomodel type of modeling.

This is a good intro to (non deep learning based) computational/synthetic biology. ML in biology is somewhat orthogonal but it is rapidly growing.

http://be150.caltech.edu/2020/content/index.html

Traditional bioinformatics is mostly text mining and string processing. Modern bioinformatics is heavily dominated by ML.