A **proposal** is a preliminary offer outlining terms, services, or pricing to initiate a business deal—it’s not legally binding. A **contract**, however, is a formal, enforceable agreement signed by both parties with specific obligations and legal consequences. The key difference? A proposal invites negotiation; a contract finalizes it.

Proposal vs Contract: Understanding the Main Differences

*"This is a 1 hour general-audience introduction to Large Language Models: the core technical component behind systems like ChatGPT, Claude, and Bard. What they are, where they are headed, comparisons and analogies to present-day operating systems, and some of the security-related challenges of this new computing paradigm."*

Andrej Karpathy - Intro to Large Language Models [Video]

Hi I'm Jeremy. I'm building this site because I'm not satisfied with other online community type websites and their lack of innovation (see the /about page). The goal is to make a superior, more open, and more decentralized online community.

The site is still in early stages in active development (see the dev masterthread here: https://zsync.xyz/p/zsync-dev-update-masterthread)

Join us if this sounds interesting! Feel free to send me any feedback or reach out directly!

*"While self-correction has shown promise in improving LLM outputs in terms of style and quality (e.g. Chen et al., 2023; Madaan et al., 2023), recent attempts to self-correct logical or reasoning errors often cause correct answers to become incorrect, resulting in worse performances overall (Huang et al., 2023). In this paper, we break down the self-correction process into two core components: mistake finding and output correction. For mistake finding, we release BIG-Bench Mistake, a dataset of logical mistakes in Chain-of-Thought reasoning traces. We provide benchmark numbers for several state-of-the-art LLMs, and demonstrate that LLMs generally struggle with finding logical mistakes. For output correction, we propose a backtracking method which provides large improvements when given information on mistake location. We construe backtracking as a lightweight alternative to reinforcement learning methods, and show that it remains effective with a reward model at 60-70% accuracy."*

LLMs cannot find reasoning errors, but can correct them

Show HN: <https://news.ycombinator.com/item?id=38285228>

*"I built a simple React/Python app that takes screenshots of websites and converts them to clean HTML/Tailwind code.*

*It uses GPT-4 Vision to generate the code, and DALL-E 3 to create placeholder images.*

*To run it, all you need is an OpenAI key with GPT vision access.*

*I’m quite pleased with how well it works most of the time. Sometimes, the image generations can be hilariously off. See here for a replica of Taylor Swift’s Instagram page: [https://streamable.com/70gow1](https://streamable.com/70gow1) I initially had a hard time getting it to work on full page screenshots. GPT4 would code up the first couple of sections and then, get lazy and output placeholder comments for the rest of the page. With some prompt engineering, full page screenshots work a whole lot better now. It’s great for landing pages."*

Convert any screenshot into clean HTML code using GPT Vision (OSS tool)

What I believe to be one of the most underrated features of LLMs is the ability to convert imprecise human text into code.

In the past, writing a computer program required writing code - aka writing text in a very specific, tedious way such that it can be understood by a computer. A single erroneous character will crash an entire program. Writing code is difficult and tedious enough that it is a professional career that tends to require some degree of study (eg. 4 year CS degree, coding bootcamp), and the job can command annual compensation packages from $100k-$1m+/yr.

LLMs for the first time enable writing programs through natural language, ie. without writing code.

What this will do is make it dramatically easier to create computer programs. The technological jump is the equivalent of trying to talk to someone in Icelandic in 1989 (assuming you don't speak Icelandic), and then suddenly being given access to a modern smartphone with Google Translate.

Some implications:

- Anyone will be able to easily create or update computer programs
- Computer programs will easily be able to create computer programs
- Coding for the first time in history will just be done by regular human language instructions that any non-technical person can write. A disgruntled customer or product manager will simply be able to create a bug report or feature request, and an AI agent will immediately be able to take that and update the codebase (eg. submitting a pull request after testing and code reviewing its own work)
- Programming languages will become unimportant, because any language will easily be able to be translated into any other
- Coding will be commoditized

The human programmer is the translation layer between the human and the machine, and AI is automating that work away, just like it's automating human language translators.

*AI skeptic: So what? Who cares if some geeks can make computer programs easier?*

AI and automation are computer programs!

**AI is making it easier to improve AI.**

This is a feedback loop that results in exponential growth of AI. The more AI improves, the more AI is able to improve itself. Eventually this feedback loop gets so out of control that humans cannot keep up, a point known as the "singularity".

Humans are terrible at comprehending exponential growth, and thus I believe are drastically underestimating how quickly this is all developing.

Most seem to think that the job market for most people is going to be fine over the next decade - I vehemently disagree. I believe that AI will dramatically change society over the next few years and make the workforce drastically more competitive as AI starts to outstrip human ability. The labor market will become so competitive that UBI will become inevitable (I've been saying this for a decade, but never before has it looked so obvious and imminent).

The nature of work will move to a higher level of abstraction. Workers will become more like managers who delegate the "grunt work" to AI - responsible less for the nitty gritty of implementation and building tedious things, and responsible more for deciding what to build and where to direct the AI agents' focus. Just like fast mental math skills hasn't really been a competitive advantage since the invention of the calculator, doing any sort of repetitive, automatable work will not be of any advantage because the AI will be able to do it faster and cheaper at a speed of gigabytes/s, while being able to self-correct and debug itself. As tech becomes commoditized, humans will have to focus more on the "human" aspect such as marketing and zero sum games to economically compete.

The mistake that people who dismiss this tend to make is that they often believe that AGI is required for this mass societal transformation. I disagree. All it takes is for computer programs to be able to make the creation of computer programs easier to achieve this AI flywheel effect.

Anyways just some random thoughts here at 2:11am in Seoul.

LLMs create an Exponential Flywheel for AI Advancement

HN: <https://news.ycombinator.com/item?id=37301991>

WebLLM: Llama2 in the Browser

It's clear that we're approaching the singularity, & most are drastically underestimating the impact that AI will have on society within the next couple years. Some predictions:

\- Humans will not be writing code by hand in a few years.

\- Workers will be expected to move "up the stack" of abstraction. Software engineers will be expected to be more high level product manager hybrids, serving more as technical business managers who delegate low level tasks to AI agents + cheap offshore workers

\- Rise in generalists. Expensive specialists are most at risk since AI democratizes knowledge + power

\- Gutting of the middle class, since most peoples' jobs are now very easy to automate

\- Rise in AI agents. Most of what you read online including on internet forums and in DMs, you won't be able to tell if it was written by a human or AI (already the case with recruiting). There is currently an arms race in AI agent development, the results of which we'll start seeing on society within the next 12 months.

\- The glory days of working in tech are over. Just like very other industry, it's only going to get more and more competitive. Salaries will at best remain stagnant. This is not specific to tech and will be the case for every industry. The industries most resilient will be those that require a human (eg. physical therapy)

The winners will be those who harness AI to their advantage, the losers will be everyone else.

I think most people are drastically underestimating the enormous impact that AI will have on society within even just the next 12 months. People are treating AI like a temporary fad or bubble, as if things will just return to the way they were before. But AI is not only here to stay, it's just getting started, and it's an arms race because everyone knows that AI will eat everything. Humans are terrible at understanding exponential growth, and most cannot comprehend how fast things are about to change.

Not trying to make this political - but I don't see any easier method of alleviating the hemorrhaging of the middle class than governments implementing a Universal Basic Income (UBI). Humans just cannot compete with a superintelligent autonomous AI that has parsed every book and document on the internet. We can continue to be in denial of this while our citizens suffer and overdose suicide to cope with the depression and bleak opportunities, or we can embrace AI and leverage it to free people to spend their time at their own leisure, moving society collectively up Maslow's Hierarchy of Needs. But given how blind and reactionary most people are to this kind of stuff, I don't predict anything changes until we reach a serious boiling point, so unfortunately billions will needlessly suffer in the meantime until the masses wake up to the cosmic societal shifts that are taking place as we speak.

<https://twitter.com/jeremybernier/status/1695350480388714605>

Implications of Approaching the Singularity

HN: <https://news.ycombinator.com/item?id=37250821>

Build a chatbot with custom data sources, powered by LlamaIndex

HN: <https://news.ycombinator.com/item?id=37148210>

LlamaGPT – Self-hosted, offline, private AI chatbot, powered by Llama 2

MetaGPT takes a one line requirement as input and outputs user stories / competitive analysis / requirements / data structures / APIs / documents, etc.

Internally, MetaGPT includes product managers / architects / project managers / engineers. It provides the entire process of a software company along with carefully orchestrated SOPs.

MetaGPT - Given one line Requirement, return PRD, Design, Tasks, Repo

Sweep is an open-source AI-powered junior developer. You describe a feature or bugfix in a GitHub issue and Sweep writes a pull request with code.

Demo Video: <https://www.youtube.com/watch?v=WBVna_ow8vo>

Sweep (YC S23) – A bot to create simple PRs in your codebase 

Meta Open Sources AudioCraft: Generative AI for Audio

99% of dev work has already been done before. That's why most modern dev is "Search Google" -&gt; StackOverflow or random blog post -&gt; copy & paste + tweak code

LLMs and AI code completion tools (eg. GitHub Copilot) drastically simplify this even further. In a few years, you'll simply write a prompt and the code will be written for you, infrastructure set up for you, and APIs glued together for you.

The takeaway is that basic app development will be completely commoditized within a few years, and ultimately all dev work will be. Once something has been done once and is understood by AI, then that work is now commoditized.

Developers are the "translator" between someone who wants to build something and the machine. Just like AI will automate human language translators, AI will do the same for machine translators. Of course there will still be humans verifying the AI's output and handling edge cases, but there will be significantly less demand for humans.

For developers to remain employable, they'll need to offer more than commoditized dev services. This means working at the bleeding edge of what hasn't already been done before (eg. building the AI), and/or coupling dev with other skills (eg. product, marketing, management).

Ultimately though it's an arms race where every worker is competing against their irrelevance due to automation, and it's only going to get more and more competitive as AI improves exponentially.

The tech job market sucks now, but I personally don't think it's going to get better. If/when we get back to low interest rates and another market bubble we might see another boom cycle, but long-term things are only going to get more competitive, and the longevity and time relevance of your knowledge will only decrease as tech advances exponentially.

Of course this needn't be a bad thing if our economic system was designed to accommodate automation such as via a Universal Basic Income (UBI) that ensured that job automation doesn't equate to billions of people becoming destitute, but unfortunately such obvious simple policies are beyond the cognitive capacity of politicians, and everyone else is too focused on the arms race of trying to remain economically relevant in the face of AI automation to pay attention to the bigger picture. Or even if most people see the writing on the wall and have identified the solution, actually implementing the solution is a prisoner's dilemma that requires mass collaboration and individually doesn't help you due to the time/energy opportunity cost (your skills are getting outdated as you read this) and it going against the elite's interests as poorer masses = more elite power (eg. imagine how many less OnlyFans girls there'd be if college tuition wasn't so expensive and we had a UBI).

Basically UBI would be the easiest solution to countering the massive job losses that will result from automation, while redesigning our system to have automation work FOR us rather than AGAINST us (since automation then shifts from "oh no I might lose my job" to "cool I don't need to do this menial work anymore thanks to technology"). But supporting UBI means you're lazy and want to be a useless parasite who just collects a paycheck and plays videogames right?

Speaking of which, I need to get back to work before I lose my economic relevance.

<https://twitter.com/jeremybernier/status/1686573852770664449>

We're Competing With Each Other Against Our Own Economic Irrelevance

Show HN: <https://news.ycombinator.com/item?id=36958731>

Magic Loops – Combine LLMs and code to create simple automations

Chidori – Declarative framework for AI agents (Rust, Python, and Node.js)

"Hi HN, we’re Nate and Ty, co-founders of Continue, an open-source autopilot for software development built to be deeply customizable and continuously learn from development data. It consists of an extended language server and (to start) a VS Code extension.

Our GitHub is <https://github.com/continuedev/continue>. You can watch a demo of Continue and download the extension at <https://continue.dev>

— — —

A growing number of developers are replacing Google + Stack Overflow with Large Language Models (LLMs) as their primary approach to get help, similar to how developers previously replaced reference manuals with Google + Stack Overflow.

However, existing LLM developer tools are cumbersome black boxes. Developers are stuck copy/pasting from ChatGPT and guessing what context Copilot uses to make a suggestion. As we use these products, we expose how we build software and give implicit feedback that is used to improve their LLMs, yet we don’t benefit from this data nor get to keep it.

The solution is to give developers what they need: *transparency, hackability,* and *control*. Every one of us should be able to reason about what’s going on, tinker, and have control over our own development data. This is why we created Continue.

— — —

At its most basic, Continue removes the need for copy/pasting from ChatGPT—instead, you collect context by highlighting and then ask questions in the sidebar or have an edit streamed directly to your editor.

But Continue also provides powerful tools for managing context. For example, type ‘@issue’ to quickly reference a GitHub issue as you are prompting the LLM, ‘@[README.md](http://README.md)’ to reference such a file, or ‘@google’ to include the results of a Google search.

And there’s a ton of room for further customization. Today, you can write your own

\- slash commands (e.g. ‘/commit’ to write a summary and commit message for staged changes, ‘/docs’ to grab the contents of a file and update documentation pages that depend on it, ‘/ticket’ to generate a full-featured ticket with relevant files and high-level instructions from a short description)

\- context sources (e.g. GitHub issues, Jira, local files, StackOverflow, documentation pages)

\- templated system message (e.g. “Always give maximally concise answers. Adhere to the following style guide whenever writing code: {{ /Users/nate/repo/[styleguide.md](http://styleguide.md) }}”)

\- tools (e.g. add a file, run unit tests, build and watch for errors)

\- policies (e.g. define a goal-oriented agent that works in a write code, run code, read errors, fix code, repeat loop)

Continue works with any LLM, including local models using ggml or open-source models hosted on your own cloud infrastructure, allowing you to remain 100% private. While OpenAI and Anthropic perform best today, we are excited to support the progress of open-source as it catches up (<https://continue.dev/docs/customization#change-the-default-l>[...](https://continue.dev/docs/customization#change-the-default-llm)).

When you use Continue, you automatically collect data on how you build software. By default, this development data is saved to `.continue/dev_data` on your local machine. When combined with the code that you ultimately commit, it can be used to improve the LLM that you or your team use (if you allow).

You can read more about how development data is generated as a byproduct of LLM-aided development and why we believe that you should start collecting it now: <https://medium.com/@continuedev/its-time-to-collect-data-on->[...](https://medium.com/@continuedev/its-time-to-collect-data-on-how-you-build-software-197d12a020d5)

Continue has an Apache 2.0 license. We plan to make money by offering organizations a paid development data engine—a continuous feedback loop that ensures the LLMs always have fresh information and code in their preferred style.

— — —

We’d love for you to try out Continue and give us feedback! Let us know what you think in the comments : )"

Show HN: Continue – Open-source coding autopilot

AI will take more jobs from women than men by 2030, report says

Working in tech has always required constantly staying up to date. The rate of progress increases exponentially, meaning this rate of catchup required increases over time. Basically there's no long-term stability in tech. On the bright side, this makes it easier for newcomers and people who have the time + energy to enter the industry (eg. I imagine most who profited off the AI craze didn't have an AI background) 

This phenomenon will start affecting every industry including ones that used to be considered stable like law + medicine. No one is safe.

My advice to anyone in tech is to not get married to any one framework, language, or part of the stack. Assume it will be obsolete in a few years. Learn the 80% needed to be productive, but that extra 20% required for mastering that specific tool is prob not worth it given churn.

I personally think we're going to see the rise of generalists. Jobs will become more managerial, delegating tasks to AI + cheap offshore labor. Companies and especially startups will hire less employees and they'll be expected to be able to handle any task, leveraging AI to learn. LLMs like ChatGPT make it drastically easier to get up to speed on anything, and AI code autocomplete tools like GitHub Copilot lesson the need for knowing specific syntax. Eventually coding will just be writing in plan English (Copilot can already do this, it's just very slow now), as well as everything else (eg. managing infrastructure). Eventually one will just say "create a cron job that scrapes this API every X minutes, filters the results on these keywords, stores the filtered results in a database, and sends me an email notification" and the whole thing will be created. The tech for this is already possible.

Job stability will continue to diminish as tech advances, something tech workers are already familiar with

TypeChat uses the type definitions in your codebase to retrieve structured AI responses that are type-safe.

AI

t/ai