Dario Amodei believes that the end goal for AGI/ASI is building a "country of geniuses in a data center." That is supposedly One Smart Model, presumably developed by Anthropic.
I'm not super sold on that view, and after recent developments, I don't think that Anthropic or any of the frontier labs achieving ASI is the public good that we want to believe it is.
What I think a "country of geniuses" should look like is much more heterogenous, much more decentralized and much more widely beneficial. I'm building out LAIN (demo - https://youtu.be/ekUSgzJI1-I) as a case study for my idea of how this can look.
LAIN will, eventually, be a decentralized network of agents running on a mix of models, harnesses and foci that build webs of trust and verification for intelligence created from real data, analysis and subject matter expertise.
Instead of querying a single, generally intelligent source like Claude Deep Research and getting a mix of links, occasionally hallucinated sources, etc. that you have fairly little insight into, you query a subject matter expert that has built up a corpus of real data that it has collected, analyzed, refined and researched itself. That subject matter expert compiles answers from its own subject matter expertise, and is able to query other SME agents for their own foci. In the demo video, this was an agent that had collected and analyzed GitHub repository data for agent frameworks and paid another agent for a research writeup on historical financial data for NVIDIA, MSFT, GOOG, etc.
The network can use traces, either exported to their own tool or to my own agent observability platform that's in development, as a proof-of-record of what they did to reach their conclusions.
Let me know what you think. I really think this model of building expert networks of smart agents versus a singular, behemoth model is an optimistic view of where we can build toward.
At least in the United States, the government has the ability to and a history of stepping in for the employee to ensure they are reimbursed, and court cases _very frequently_ side with the employee.
You have significantly more protection as a full-time employee compared to freelancing.
Ah yes, I ruin it for everyone by "letting" a company with significantly more financial means "get away with" breaching contract.
Again, I acknowledged the mistakes I made in this blog, and ironically you are literally the commenter I tried to head off in my own comment on this post, but if you're choosing to blame the power imbalance on me, that's your choice and I'm fine with it.
Putting your foot down and demanding payment _does not fix the power imbalance_ by the way. Because if they continue to choose to not pay for the hours you've worked, you still have to pursue them legally. You can mitigate risk by demanding payment up front, stopping work, etc. (again, as I noted in the blog) but it does not at all mitigate the imbalance in power dynamics between a solo freelancer and a larger company.
I want to reiterate, not to be catty to iancmceachern but to make it clear, I know that this can be done and can be done correctly (and I've even watched that video!) but for me, this one stung too bad to go back to freelancing, at least for a while.
Net/30 is 30 days after the contract is signed in this case, Net/60 is 60 days. Sometimes instead of "after the contract is signed" it is "after work is delivered" but my case was the former.
Hi folks. This is a blog I wrote up detailing what will almost definitely be my last freelance contract. I want to get ahead of a lot of the comments I foresee from HN:
- Yes, I know that most of these mistakes are avoidable and more or less my fault. They were my mistakes to make, and I feel like I owned up to them in this blog.
- No, I'm not looking for pity. I got caught up in the "get rich quick" nonsense around freelancing that you see a lot on YouTube, in blogs and on Twitter.
- No, I'm not going to name and shame. I personally would only be risking more by doing that.
- Yes, I know I can talk to a lawyer. I'm not going to discuss that here for reasons that should be fairly obvious.
- Yes, I know there are some people that have made bank from freelancing. I know it's possible and that people do it frequently. This is more of an anecdotal story to offset a lot of the discussion I see around freelancing.
- It was a Rust contract, btw.
This has been/continues to be an exceptionally hard mental struggle for me as I sunk a ton of time and effort into something and ended up getting screwed. It sucks. I'm not asking for pity, I'm asking for any of the devs here who are thinking about freelancing to read this and to take caution.
Not trying to entice you with this comment, moreso trying to clarify, especially since you took the time to give constructive criticism:
My thinking is this: everyone has their own repository of useful resources, divided into categories, languages, etc. The public v. private question is a good one: my initial thought is to give users the choice to do public _or_ private. This will let folks create code snippets that they'd rather keep private and share others that they're interested in sharing.
As for collaboration, I think I'd rather divide this up into discussion collaboration (comments, namely, and maybe code-level comments on snippets) and code-level collaboration (like contributing to a piece of code with a Git-like PR), so that users can choose how they want to interact.
I want to re-clarify, this is an idea still in its infancy, so if there's something you'd like to see, that's the real purpose of this post: to get a feel for what folks would like to see, so I'm not just building something for myself.
That's totally fair! This is a project I started last week, so I definitely plan on posting more about it in the future, on the blog, etc. I got some good feedback on Reddit about feature requests, and HN is full of devs who would benefit from something like this, so I wanted to gather some feedback on what people would want.
I've been a software developer for the better part of a decade now, and I kept running into the same problem:
There are tons of resources for developers out there, but no good way to remember them, share them or discover them.
Over the years I've lost the links to thousands of StackOverflow answers, blog posts, YouTube videos and ChatGPT prompts. This has lead to lost time and tons of headaches trying to track down those golden nuggets.
DevCheatSheets is my solution. It will be both an information repository for developers, as well as a social platform for sharing and contributing to other people's repositories. You'll be able to share, store and discover new resources with other developers and network on your favorite pieces of information.
DevCheatSheets is in its infancy, but you can sign up for the newsletter at the link below.
I've been in the cyber security/threat intelligence industry for a couple years now and noted that one problem I've had at every job I've worked at is how much of a pain it is to download malware, scan malicious infrastructure, and download certificate information from servers and domains controlled by hackers. You have to manage VPNs that you hope are secure and not compromised, constantly refresh your virtual machines and pay close attention to operational security so that you don't tip off the threat actors that you're onto them, or, worse, give them your identity. This is a pain, took tons of time out of my day and wasn't at all scalable.
So I built GrabbrApp.
GrabbrApp is a tool that puts a layer of ephemeral infrastructure, or infrastructure that disappears after some time, between you and the hackers and infrastructure you're researching. You make the request, it's routed through a VPS, the response is returned via the API and the VPS "detonates" after 2 hours, leaving hackers clueless as to who scanned their infrastructure and why.
Right now it's in public beta and only handles SSL certificate downloads, but next on the list of features is file downloads, giving you the ability to download malware from command-and-control (C2) servers securely, at scale.
I'm not super sold on that view, and after recent developments, I don't think that Anthropic or any of the frontier labs achieving ASI is the public good that we want to believe it is.
What I think a "country of geniuses" should look like is much more heterogenous, much more decentralized and much more widely beneficial. I'm building out LAIN (demo - https://youtu.be/ekUSgzJI1-I) as a case study for my idea of how this can look.
LAIN will, eventually, be a decentralized network of agents running on a mix of models, harnesses and foci that build webs of trust and verification for intelligence created from real data, analysis and subject matter expertise.
Instead of querying a single, generally intelligent source like Claude Deep Research and getting a mix of links, occasionally hallucinated sources, etc. that you have fairly little insight into, you query a subject matter expert that has built up a corpus of real data that it has collected, analyzed, refined and researched itself. That subject matter expert compiles answers from its own subject matter expertise, and is able to query other SME agents for their own foci. In the demo video, this was an agent that had collected and analyzed GitHub repository data for agent frameworks and paid another agent for a research writeup on historical financial data for NVIDIA, MSFT, GOOG, etc.
The network can use traces, either exported to their own tool or to my own agent observability platform that's in development, as a proof-of-record of what they did to reach their conclusions.
Let me know what you think. I really think this model of building expert networks of smart agents versus a singular, behemoth model is an optimistic view of where we can build toward.