Logisite🖲️

@doom_scroller__ @evisdrenova Opencode user Haven't compared it's compaction with another harness. Have you compared pi with another?

@Logisitecr @evisdrenova Its unavoidable. Not all compaction strategies are equal either, - pi-coding-agent does a bang up job I've noticed.

my ideal AI interface is a single never-ending chat thread. i don't want to think about the concepts of sessions, context windows, worktrees, mcp servers or really anything else. The harness should automate everything transparently.

@iamumairamjad @rxhit05 Yes in that case of a defined product. But from day one of learning SWEing, maybe not

@Logisitecr @rxhit05 When someone knows what, why and for which audience he is making product. He should build the Social media foundation around them from day 1.

hot take: the average developer spends more time learning than building. that's why they never make money.

@esteexbt 10

I just got a new hair cut, rate /10

@esteexbt @Bruce1Offordile @genesystechhub Yayy

TASK 6 done and dusted Learnt how to use Fig Jam while at it Isn't that interesting? Mehnnn Fig Jam is dope

@esteexbt Anticipating

Putting this out here, because after the month of July. I will be fully monitized on X.

@Av1dlive Nice and big How much tokens and context does it take to transverse this?

You can build an AI second brain in 15 minutes. No coding experience needed. no $1000 course [Here is how you can do it in 5 mins:] Step 1: Download Claude Desktop. Step 2: Download Obsidian Desktop. Step 3: Create a new vault and start dropping .MD files into it. Step 4: Tell Claude Code to connect directly to your vault using Andrej Karpathy's prompt: gist.github.com/karpathy/442a6… That is it. Your entire knowledge base becomes searchable, connectable, and queryable by the most powerful AI model on earth. Every note you have ever written. Every idea you have ever captured. Every resource you have ever saved. Claude can now read all of it, find connections you missed, and surface insights from your own thinking that you forgot you had. Most people are using Claude as a search engine. The people building second brains with it are using it as an intelligence layer on top of everything they know. The gap between those two use cases is the gap between asking Google a question and having a research partner who has read everything you have ever written. Bookmark this. Build it tonight.

@yourmediaguy_ 🟢

We got a new look don’t we? Yes we did💜

@Iamceejaymac Why are these odds this high. When Messi to score happens a lot?

Messi came off the bench. 20 minutes later... Free-kick goal. 🐐 I had a live bet on Messi to score or assist, so that strike turned my ticket green. Moments like these remind me how much fun player markets can be during live games. One thing I'd love to see more prediction market platforms explore is player-based markets: • Score or assist • Shots on target • Assists • Cards They'd make live football even more interesting. What's your favorite player market to trade?

@udohjeremiah_ How about you bring some quotes that failed 🙂

Before you dismiss AGI, remember: “I think there is a world market for maybe five computers.” — Thomas J. Watson , IBM chairman (1943) “There’s no reason for individuals to have a computer in their home.” — Ken Olsen, DEC founder (1977) “640K bytes of memory ought to be enough for anybody.” — Bill Gates, 1981 “I predict the Internet will soon go spectacularly supernova and in 1996 catastrophically collapse.” — Robert Metcalfe, inventor of Ethernet (1995) “People don’t want to watch TV on a tiny screen.” — early skepticism around YouTube/mobile video

@omopeweb3 Gm

GM $MON 💜 & $SUI 🩵 CT Happy Sunday

@udohjeremiah_ Fair enough

@Logisitecr I wouldn’t want to call out anyone in particular (don’t think it’s fair), but if you look around you’ll find them.

Most Nigerian tech summits confuse me. You call it a “tech summit” to showcase the future of technology, but then invite mostly entertainment-focused creators as speakers (because they're popular). Nothing wrong with creators, but how does that align with the goal of a tech summit?

@davidokocha086 Given what you are optimising for (coverage) that tradeoff works. About your phase 5, where you'll add a whisper model. Do you intend to load and unload it on request to save baseline compute costs ?

llama.cpp already puts pressure on low end hardware, adding an embedding model at query time on top of that is not feasible. I’d rather precompute the vectors once and ship them with the curriculum db. FTS5 adds zero memory overhead and runs inside SQLite, no extra process. knowledge graphs need the curriculum fully mapped first and that comes later.

shipped phase 4 of the EduOS ai engine stripped python out completely. model loads from a GGUF file, inference runs in C — no ollama, no subprocess, no cloud. the AI now works fully offline on a 940MB file. still cleaning up the RAG but the pipeline is live. offline AI education on low-end hardware is becoming real. github link 🔗 below

@SirGparody Like what?

being broke and sick at the same time is a bad combo. everything will be looking like photosynthesis 😹

@Romeo_Onchain Yeah

Stay glued to Ansem..when he says he got it he actually does

@udohjeremiah_ Important? yes. Biggest? Can't know

Cold take: The biggest skill of the next decade might be attention.

@MystiqueMide Soon

A moving man will meet his first hackathon win one day ,i believe 🙏🏾

@Romeo_Onchain Dad

being a supportive man as usual Support our new business...come have a taste of AMAKA LUXURY AKARA

@EOEboh Default more network and stuff

Being in tech means constant passive learning.. you may be in the shower and still asking yourself "what's really a double in java"?

@NitinthisSide_ You literally just said the answer in the question.

My colleague interviewed an AI Intern candidate today. He asked: “What is an AI Agent?” The candidate replied: “An AI Agent is a system that can reason, plan, and use tools to accomplish tasks autonomously.” A good answer. Then came the follow-up: “If an LLM can answer questions on its own, why do we need AI Agents? When would you build an agent instead of just calling an LLM API?” That’s where the conversation got interesting. Knowing AI terminology is one thing. Knowing when to use an LLM, RAG, fine-tuning, workflows, or agents is what makes someone an AI engineer—not just someone who has built a chatbot. So here’s the same question for you: When would you choose an AI Agent over a simple LLM API call?

@iamumairamjad @rxhit05 That's too early. The preferred approach is to know when you've learnt enough. Which is ofcourse hard to know

@rxhit05 My take: he should focus on distribution from day 1 to make money