hexman

5 ms·

1.
by hexman 2y ago
Seems like an interesting idea to abandon the scaffolding code to create agents. Thoughts?
2.
Agents in Markdown syntax with loops, memory and tools(github.com)
1 points by hexman 2y ago|1 comments
3.
Show HN: Cognitive Language Agents (Executive Functions)(github.com)
4 points by hexman 2y ago|0 comments
4.
Show HN: Mentals – How I Implemented CPU Logic into LLMs (Open-Source)(github.com)
3 points by hexman 2y ago|1 comments
5.
by hexman 2y ago
For the last four months I have been experimenting a lot with LLM models in an attempt to simplify the creation of reasoning chains without using an auxiliary programming language. The idea comes from the control unit in classical processor
6.
Are LLM models approaching a saturation point?(x.com)
1 points by hexman 2y ago|1 comments
7.
by hexman 2y ago
LLM benchmarks are normalized between 0 and 100. The main benchmarks alredy close to 100: - common sense reasoning (WinoGrande) - arithmetic (GSM8K) - multitasking (MMLU) - sentence completion (HellaSwag) - common sense reasoning 'chal
8.
by hexman 2y ago
Just implemented on gpt-4o a team of 4 agents that was able to build a 2d platformer. The only model that managed to accomplish the task. 4 agents: - Root agent: task description, Game Design Document generation. - Project manager based on
9.
Mario like 2D platformer game AI generator(x.com)
2 points by hexman 2y ago|1 comments
10.
10M Context Length LLM Inference | UC Berkeley(arxiv.org)
4 points by hexman 3y ago|0 comments
11.
by hexman 3y ago
Imagine if Apple's or Google's cloud went down and all your apps on iPhone and Android were broken and unavailable. Absolutely all apps on billions of phones. Cloud =! OpenAI Clouds store and process shareable information that mul
12.
LLMs are new CPUs, not Operation Systems(lesswrong.com)
1 points by hexman 3y ago|0 comments
13.
by hexman 3y ago
Asked GTP4-32k to make an article summary of 10 important points. And then applied the summary to the software architecture --- Public and Shared Nature of Software Architecture: Just like physical architecture, software architecture is inh
14.
by hexman 3y ago
I found that the approach of template processing at large prompts leads to difficulty in reading programs. Their attractive part is that control flow is not separate from prompt as in langchain, which allows you to write prompts as classica
15.
Home Depot with 2.3K edge locations on k3s like Chick-fil-a(datacenterknowledge.com)
3 points by hexman 4y ago|0 comments
16.
Turing RK1 modules performance and ETA(kickstarter.com)
1 points by hexman 4y ago|0 comments
17.
by hexman 4y ago
Pirated content what do you mean?
18.
by hexman 5y ago
I'm shocked. I have the same mindset. For example, if I need to predict how technologies can potentially evolve, I visualize them in different planes based on the history of technologies, such as centralization/decentralization cy
19.
by hexman 5y ago
idk bad in wordceling
20.
by hexman 5y ago
Yes :)
21.
by hexman 5y ago
Awesome!
22.
by hexman 5y ago
10 shapes 12 sec/shape (2nd try, no brute-force, mobile) 16 shapes (3rd try) 18 shapes (4th try) 24 shapes (5th try, no brute-force) P.S. I had 7 years of art school in addition to engineering
23.
by hexman 5y ago
Maybe it's time for a class-action lawsuit?
24.
Geekbench RK3588 vs. Raspberry Pi 4(browser.geekbench.com)
3 points by hexman 5y ago|0 comments
25.
by hexman 5y ago
Too early
26.
by hexman 5y ago
* ipfs peers and cluster peers * ipfs swarms * Ethereum transaction and mining nodes * Ethereum 2 beacon and validation nodes * Private Ethereum networks * Filecoin nodes * Filecoin backed pinning services (FPS) * Polkadot * Algorand
27.
Kubernetes Blockchain Operator(github.com)
2 points by hexman 5y ago|1 comments
28.
by hexman 5y ago
... DilDog is this you?
29.
Planet-scale Kubernetes(liqo.io)
1 points by hexman 5y ago|0 comments
30.
by hexman 5y ago
Hey What does the hacker community think about this? pondered https://turingpi.com/turing-pi-v2-is-here/

More ›