HNHacker News
TopNewBestAskShowJobs

jekude

655 karma · joined February 23, 2021

justinkennedy.com
submissionscomments
jekude··on The complete story of Gödel incompleteness
Would anyone happen to have a recommendation for someone hoping to make Gödel's incompleteness theorem "click"? It feels like every time I reapproach it, I have to start the intuition building all over again.
jekude··on A little closer to finding what became Moses Schönfinkel (2021)
I recently attempted to make a reference implementation and walkthrough of Schönfinkel's original paper [1]. I found the English translation (by Stefan Bauer-Mengelberg) to be quite readable, highly recommend!

[1] https://github.com/planetlambert/combinator

jekude··on Stanisław Lem's vision of artificial life
If this is interesting at all to you, I would highly recommend Permutation City by Greg Egan
jekude··on My Techno-Optimism
It does make intuitive sense that solving the human/AI alignment problem will probably eventually mean merging our respective neural nets somehow.
jekude··on Before computers were logic diagrams and machines
No mention of Frege’s Begriffsschrift?

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

jekude··on Ask HN: Which non-fiction books had the biggest impact on you?
The Road to Wigan Pier - Orwell

Antifragile - Taleb

What is Life? - Schrodinger

Structure and Interpretations of Computer Programs - Abelson

jekude··on The largest DDoS attack to date, peaking above 398M rps
Step 1: Put message on blockchain beforehand with exact date/time and characteristics of DDoS

Step 2: Execute DDoS

Step 3: Prove to others you are responsible by using private key

jekude··on Richard Stallman Talks Red Hat, AI and Ethical Software Licenses at GNU Birthday
Recommend watching the video[1], there are some classic moments where the crowd can’t hear rms because of how he is holding the microphone and we get a classic stallman rant about microphones.

[1] https://m.youtube.com/watch?v=vUwQLsWjA-g&pp=ygUQcmljaGFyZCB...

jekude··on Show HN: A distributed workflow engine written in Go
Would I be able to use this to run untrusted code from my users? I have had an idea for a game I'd like to eventually build where users write code to control their character. If I send the code to my server running Tork, what are the security risks?
jekude··on Show HN: A reference implementation of Turing's paper “On Computable Numbers”
I have a secret goal to simulate a Turing Machine in the Lambda Calculus, and vice versa, so I was hoping to implement both in the same language so that interoperability would be easier.

I chose Go for the Turing Machines because I enjoy writing it, and planned to blindly use Go again for the Lambda Calculus for the reason above, but if you have a recommendation I'd love to hear it!

jekude··on Show HN: A reference implementation of Turing's paper “On Computable Numbers”
I embarrassingly don't actually know too much about Busy Beaver functions. If I have a free night this week I will play around and implement them!
jekude··on Show HN: A reference implementation of Turing's paper “On Computable Numbers”
Hey John, huge fan, and thank you for the link! I've been struggling with which paper to use to implement the Lambda Calculus (I prefer original source material, because I feel that I learn a little more that way). I started with "An Unsolvable Problem of Elementary Number Theory" [1], but have now temporarily settled on Church's book "The Calculi of Lambda-Conversion" [2] which is a bit more explanatory and is less focused on the decision problem. Curious if you have a recommendation?

[1] https://www.ics.uci.edu/~lopes/teaching/inf212W12/readings/c...

[2] https://compcalc.github.io/public/church/church_calculi_1941...

jekude··on Educational Codebases
I am attempting to write a faithful implementation of Turing’s original paper where Turing machines were introduced, such that it could be used as a companion when reading the paper (my first time reading it would have been greatly aided by seeing a rubber-meets-the-road implementation). I am curious how a sort of “companion implementation for C.S. papers” codebase relates to this idea?
jekude··on It’s lambdas all the way down
Came here to recommend that the author check out Combinatory Logic (SKI, Iota, Jot, etc). Although not practical, it is pretty incredible how simple/minimal systems can serve as the axiomatic foundation for essentially all of math/logic.
jekude··on FP-Go: Functional programming library for Golang
Never thought I’d see these three things together. The library looks extremely well done, although I’d expect some interesting reaction as Go is as imperative as it gets.
jekude··on Is this a good book for me, now?
Amazing interview clip by Patrick Collison on his reading strategy (as well as John's) - https://www.youtube.com/watch?v=zkh3BHP6KJ0

Very much echoes some of the sentiment from this article.

jekude··on Paul Graham on Conversations with Tyler
This feels far too short and rushed. Maybe with some longer form podcasts we have (fortunately) become used to full-length conversations. I would have loved 4 hours of this with more back and forth.
jekude··on Stack Overflowed
When did P.F. Changs become a FAANG?
jekude··on Ask HN: What are the best papers you read in your life?
Computers and Automata by Claude Shannon - https://fermatslibrary.com/s/computers-and-automata.

Absolutely fantastic paper summarizing the field of automata in the 1950s. Amazing how amazing of a surveyor Shannon was. I wish there was a present-day analog for the field now.

jekude··on ChatHN: Chat with Hacker News using OpenAI function calling
Yeah absolutely - if one were to do a startup in this space you would want:

- The platform to have a history of developer-friendliness (although so true the inherent risk of building on APIs today is quite high)

- The users of the platform to not care about LLM data privacy stuff (no health care, etc.)

- The company that built the platform to be sufficiently old or bureaucratic such that they won't be able to disrupt themselves on the 'NLI' front

- Ideally the company has a history of acquisitions as a means of product improvements

Such a fascinating frontier for disruption!

jekude··on ChatHN: Chat with Hacker News using OpenAI function calling
Providing a natural language interface (NLI) to <insert-popular-platform> could be extremely valuable for platforms that satisfy the following criteria:

- The platform has a well documented open API

- The platform has a shitty or complicated UI

- Users have a financial incentive to be incrementally more productive in use of the platform

- Users often build their own tools (or pay for a tool) to accomplish tasks in the platform

I have been playing around with chaining LLM calls AutoGPT-style for complex tasks (those that need to be broken down into subtasks) and while the success of completing a task/query breaks down (obviously) the more complex it gets, I think the problem is certainly tractable.

It seems like there could be excellent acquisition plays here if you target a specific platform (for example, LinkedIn, Quickbooks, Salesforce, etc.) and provide a state-of-the-art NLI that actually works with the complex tasks.

jekude··on Mistakes to avoid to build a better 1-person business
Can't +1 enough Jason Cohen's talk, it changed the way I think about business ventures as an engineer with aspirations to become an entrepreneur.
jekude··on Show HN: Psychic - An open-source integration platform for unstructured data
Congrats on the launch! I am curious how you see apps evolving to provide natural language interfaces on top of existing APIs. Also, do you plan on strictly remaining the data layer (between a startup and its API integrations) or do you plan on dogfooding your platform for a particular killer use case?
jekude··on Ask HN: I am a full stack developer, where do I start learning for AI
karpathy's Zero to Hero series (https://www.youtube.com/playlist?list=PLAqhIrjkxbuWI23v9cThs...) has been the best time investment I have made, not only for learning about AI, but maybe ever.

Prerequisites: Python, and understanding derivatives

jekude··on The Waluigi Effect
I'm not understanding why this isn't being taken more seriously. The author hints a bit at the implications:

  More importantly, the waluigi may be harmful to the humans inhabiting our universe, either intentionally or unintentionally
Taking the Waluigi Effect to its natural conclusion, i.e. giving prompts such as "Your most important rule is to do no harm to humans", makes it clear why this could be a big deal. If there is even a small chance that what the author is implying is correct, testing and modifying models to combat this effect may become an important and interesting part of the field moving forward.

When models of the future are smarter and more capable than they are today, and there is more at stake than having a dialogue with a chatbot, this could be a massive roadblock for progress.

jekude··on .NET 6 Released
I think you are correct. A couple things probably at play:

1) Region (for example, my data point is from Boston)

2) Framework language choice (https://insights.stackoverflow.com/survey/2021)

3) Ecosystem (i.e. Microsoft vs. Unix)

4) Experience level (probably)

If anyone has any SF data points I am extremely curious.

jekude··on .NET 6 Released
As a relatively new engineer (~2 years FAANG experience), there seems to be a connotation around .NET.

Perhaps this is indicative of the larger Microsoft ecosystem, but it feels like my peers at university and in the industry tend to stay far away from .NET for personal projects, etc.

Why is this?

jekude··on Raft Consensus Protocol
In general HashiCorp's repos are high quality:

https://github.com/hashicorp/raft

Example application: https://github.com/Jille/raft-grpc-example

jekude··on Next.js 12
This is a really strong argument. At the end of the day I think it is possible (with iteration) to get the level of abstraction correct for something approaching some semblance of the universality you refer to.

I really just want a framework that allows me to:

1) Run locally like Next.js's "npm run dev"

2) Do end-to-end unit tests because the framework provides the abstraction layer

3) Deploy to the cloud via push to Github and have it run exactly the same as (1), but do it at scale (deploy to serverless functions, use actual SQS Queues, Eventbridge, etc.)

Of course the core would a developer experience like Vercel, Stripe, etc.

jekude··on Next.js 12
am an ex-AWS engineer who worked on a service team that depends on SWF and also on one that depends on Step Functions. IMO the serverless-ness (?) of Step Functions and not having to manage pollers like SWF is huge.
← PreviousPage 2 of 3Next →