T!

Software Engineer

[T! @unkletayor]

"I would rather live my life as if there is a God and die to find out there isn't, than live as if there isn't and to die to find out that there is".

- Albert Camus

[Oluyomi Ojo @OluyomiOjo]

Talking about being in rooms that change your trajectory.

The Internet Society Fellowship Program is one of them. This 10-month program enables fellows to turn locally grounded ideas into tangible results and lead lasting change.

Fellows will choose from three specialized tracks—Internet Policy and Mobilization, Internet Connectivity, and Internet Safety and Security—and are eligible for both travel support to relevant events and a stipend to cover direct project costs.

Enrollment for the Pre-Fellowship Training is open until 22 November. Don't sleep on this one.

https://t.co/TMKhYBnN00

[👑S.A.L.A.K.O🕊 @UnkleAyo]

Dear men,

It impresses upon my soul to reshare my thoughts.

I've read so much about Diogenes but one quote of his that massively influenced my value system, reads:

"It is the privilege of the gods to want nothing, and of godlike men to want little".

It imprinted upon my soul for a long time, especially as someone that's ambitious.

Friends, verily I say unto you:

Ambition is good but nothing is more frustrating to a man that for him to exist at the edge of a precipice where there are more factors out of his control, than within his control.

Endless chase in ambition without a defined scope is outside your control & with every season you have a new high to chase without a moment to enjoy where you are.

It becomes easy to discard your present as a valley as you glare at a new high, forgetting that your current valley used to be a mountaintop you aspired to.

I guess the point I'm trying to make is, have a number in mind. Especially as an ambitious person, have a number in mind.

The endless pursuit of ambition will leave you with a hollow soul. It will strip you of the beauty of momentary happiness, satisfaction of wins & incredible undervaluation of how far you've come.

As a young hustling man, it is important that you define what happiness and success looks like for you, and to do so early on.

The ultimate question now becomes - Are you in control of your dreams? Or are your dreams in control of you?

The answer to this question is a predictor of your future happiness or lack thereof.

I hope my writing found you well.

As you were.

[Oreoluwa Bukola, CFA. @BukkyOA]

This is my interpretation of life:

No matter how important you are, life is bigger than you. So, life continues with or without you.

Life does not wait for you.

Whatever your contribution, it is in service of your purpose ONLY. Others may benefit as free riders, but you are not doing anybody favours by chasing your dreams.

So, if you think that people will drop what’s in their hands to sit still because of you, it will not happen.

And that is why you must live your life to the fullest ON YOUR TERMS.

[» teej @teej_m]

Internet traffic grew 100x between ‘98 and ‘03. Every interesting website was invented in this era: Google, Wikipedia, MySpace, Wordpress, Livejournal, 4chan, Newgrounds, Craigslist, etc.

To make an interesting website, you need shared state. The only way to get shared state is with a database. That meant you used MySQL.

MySQL is a program that reads and writes shared state to disk. Disk drives were slow and expensive. A modern microSD card the size of your fingernail is 50x better and 10x cheaper than a disk drive in 2000. We used 32 bits back in those days, so a single machine capped out at 4GB of memory. MySQL didn’t support sharding natively, so you’re either stuck to a single machine or you write your own query router.

So your database, the thing that’s central to making your website cool, is constantly overloaded. Your website is going down and everyone is leaving. You’re barely treading water while the internet is growing 2% a week.

Then, the absolute legend @bradfitz, trying to keep LiveJournal online under 1000 req/sec of traffic, invented memcached. As the story goes, he was waiting for the shipment of a new physical server for his database when he realized his application servers had lots of unused memory. He could turn this into a shared network cache that would drastically reduce load on his database.

This new piece of software completely changed how websites were made. Within a few years, memcached was a mandatory part of the web stack, you couldn’t scale without it.

Some time later, @antirez came along and made a modern replacement with a lot of really nice features called Redis.

[GREG ISENBERG @gregisenberg]

This is the MOST asymmetric window I've ever seen in business.

It's honestly hard to comprehend. Never in history have this many categories opened at once, ALL ripe for the taking:

1. Take a boring business doing 20% margins, an accounting firm, an IT shop, a local service company, and swap the human labor for agents until it runs at software margins with customers it already has.

This is the Thrive Holdings model.

2. Build the picks and shovels, the clean API or data source that agents call 1000 times a day, and get paid per call instead of per user.

I call this being in the building blocks business.

3. Buy a sleepy SaaS with real customers and a tired founder, then rebuild it AI native and 10x what it can do while building a media machine along side it.

4. Run a service business that looks like an agency on the outside but is really you plus a team of agents doing the delivery, priced like consulting, delivered like software.

5. Watch what people are still hiring for on Upwork, the "data entry specialist," the "appointment setter," the "research assistant," and build the productized agent that replaces that exact role, then sell it to every company posting the job. Vertical agents are here!

I could keep going forever but you get the point.

Contrary to what a lot of people believe (AI is coming for us, you need to raise venture to compete etc), the opportunity right now is endless.

You don't need to quit your job. Finally, you can build a real startup on the side.

It's starting season.

If you've EVER thought about starting a something of your own, this is the moment to stop thinking about it.

Seasons like this don't last forever.

Start.

[T! @unkletayor]

@abrahamtayo @instablog9ja https://t.co/94gOYvHEj3

[Crazy Codes🚀👨‍💻 @richsongocrazy]

I just cracked a new way of getting and applying for jobs at startups that just got founded on LinkedIn if you want to get jobs through this, check this thread 🧵

[Suraj Sharma @suraj_sharma14]

If you can build these 12 Agentic AI projects.

You're hired.

Project 1: Structured Output Agent
Enforce Pydantic JSON schemas, validate tool responses, retry on parse errors, log validation failures.
→ Shows: You can make LLMs reliable not random

Project 2: RAG Agent with Citation Grounding
Retrieve context, generate answers with sources, flag low-confidence responses, fallback to search.
→ Shows: You can prevent hallucinations at scale

Project 3: ReAct Planning Agent
Observe → think → act → reflect loop, max iteration limits, self-critique, graceful degradation.
→ Shows: You can build agents that don't infinite loop

Project 4: Multi-Tool Orchestrator Agent
Dynamic tool registry, capability-based routing, permission scoping, parallel execution, conflict resolution.
→ Shows: You can coordinate complex workflows

Project 5: Memory-Enabled Conversational Agent
Short-term buffer + long-term vector recall, context compression, relevance scoring, cross-session sync.
→ Shows: You can build agents that remember users

Project 6: Human-in-the-Loop Approval Agent
Uncertainty detection → pause → request human input → resume with validated context, full audit trail.
→ Shows: You can build safe, compliant systems

Project 7: Cost-Aware Agent Router
Token budgeting per task, model routing by complexity/cost, early exit on confidence, cost-per-decision analytics.
→ Shows: You can reduce infra costs by 40-60%

Project 8: Event-Triggered Automation Agent
Listen to webhooks/queues, execute workflows on triggers, idempotent execution, dead-letter handling, retry logic.
→ Shows: You can build production automation not demos

Project 9: Multi-Agent Debate System
Multiple agents propose solutions, critic evaluates, voting/consensus logic, aggregator synthesizes with confidence.
→ Shows: You can orchestrate swarms not single agents

Project 10: Self-Reflective Agent with Auto-Eval
Execute → evaluate via LLM-as-judge → critique reasoning → regenerate with constraints, log improvement metrics.
→ Shows: You can build systems that improve over time

Project 11: Production Agent with Observability
Deploy with LangSmith/Arize tracing, latency/cost dashboards, alerting on loops/failures, canary testing, rollback.
→ Shows: You can ship to production not just localhost

Project 12: Open Source Agent Framework Contribution
Extend LangGraph/CrewAI/AutoGen with new pattern, write docs + demo, publish benchmarks, submit PR + tutorial.
→ Shows: You're a community builder not just a consumer

(Bookmark this)

happy building...

[paperpaper @paperpaper886]

最近在带入组的本科实习生，发现怎么读论文其实是科研训练里最容易被忽略的一步。
推荐一篇每个科研新人都该读的经典短文：S. Keshav 的 How to Read a Paper。

文章提出了非常实用的“三遍读论文法”：
第一遍，5 到 10 分钟快速扫读：标题、摘要、引言、章节标题、结论和参考文献。
目标是回答 5C：
Category, Context, Correctness, Contributions, Clarity。
也就是判断这篇论文是什么、和谁相关、假设是否合理、贡献是什么、写得清不清楚。
第二遍，认真读论文主线，但先跳过证明细节。重点看图表、实验设置、结果是否清楚、引用了哪些关键工作。
第三遍才进入深度理解：尝试像复现一样重建作者的思路，检查假设、方法、创新点和潜在漏洞。

放在今天看，这个方法和 AI 辅助读论文其实很契合。
第一遍可以让 AI 帮忙快速总结论文的研究问题、核心贡献和主要结论，但自己一定要判断这篇文章是否真的值得继续读。
第二遍可以让 AI 帮忙解释方法、实验设置、图表和不熟悉的概念，但不能只看 AI 总结。关键图表、实验设计和结果数字一定要回到原文核对。
第三遍可以让 AI 扮演 reviewer，帮你追问：这篇文章的假设是否成立？实验是否支持结论？有没有 missing baseline？有没有潜在的数据泄漏、评价偏差或过度 claim？

读论文不是“读完”就行。真正重要的是知道什么时候快速跳过，什么时候认真理解。
尤其在 AI 工具越来越强的情况下，科研新人更需要训练自己的判断力。
AI 可以帮你压缩信息，但不能替你决定一篇论文是否重要、是否可信、是否值得借鉴。

https://t.co/8gUc4HbLwR

[The Grid Post @TheGridPost]

BIG CHANGE:
YOU CAN NOW OFFICIALLY
GENERATE ELECTRICITY FOR YOURSELF, USE WHAT YOU NEED, AND SELL EXCESS ELECTRICITY BACK TO THE GRID THROUGH YOUR DISCO

The Nigerian Electricity Regulatory Commission (NERC) has commenced the Net Billing Regulations 2026, a new framework that allows eligible electricity users to generate power mainly through solar energy for their own use and export excess electricity back to the distribution network.

In simple terms, if your solar system generates more electricity than you consume, you can now send the extra power back to your DisCo and receive credits under a regulated billing arrangement.

This effectively creates what the sector calls a “Prosumer” ;meaning you are both a consumer and producer of electricity.

However, there are conditions.

This is not yet targeted at the average small residential solar setup.

To qualify:

-You must already be connected to a DisCo network.

-Your renewable energy system must have a minimum installed capacity of 50kWp and a maximum of 1.5MWp

-You must obtain approval from your DisCo.

-You must sign a Net Billing Agreement and register with NERC.

Approved users will receive bidirectional meters that track: ➡️Electricity imported from the grid
⬅️Electricity exported back to the grid

[Phuong Le @func25]

Chapter 8 of The Anatomy of Go is now available. It discusses channels, timers, concurrency, goroutine creation and lifecycle, runtime startup, preemption, scheduling, I/O handling, the system monitor... and a bunch of other things.

The book is now complete at 873 pages, and early access will be closing soon after the editing phase, (which may add a few more pages).

As usual, here is an excerpt:

[Logan Thorneloe @loganthorneloe]

Read this to get started learning ML infra.

This is an excellent high-level overview of important considerations in ML training from CMU. It touches on:

- hardware
- memory
- the ML experimentation process

https://t.co/RTWm0Ecni1

[Ahmad @TheAhmadOsman]

Step-By-Step LLM Engineering Projects Roadmap

- Build a tokenizer
- Learn embeddings
- Implement RoPE / ALiBi
- Hand-wire attention
- Build MHA
- Build a Transformer block
- Train a mini-former
- Compare objectives
- Build sampling
- Speculative decoding
- KV cache
- MQA / GQA / MLA
- Long context
- FlashAttention
- Hardware budgets
- Toy MoE
- Sparse model trade-offs
- State-space / linear attention
- Diffusion language models
- Data pipelines
- Synthetic data
- Scaling laws
- SFT / DPO / RLHF / GRPO
- Quantization
- Serving stacks
- Eval harnesses
- RAG
- Tool use / agents
- Vision-language adapters
- Interpretability
- Red-team suite
- Full capstone model system

One request:

Choose an Opensource AI lab when you make it

Opensource is where humanity gets to keep the tools

DM me when you've made it ;)

[Idowu Anthony-Ajileye @dobmarch8]

Almost 13m units bid for TIP right now and not a single offer. This is when a single news drives investment decision rather than conviction based on fundamentals and progressive trajectory of an Organisation.

How l wish Investors read because majority don't and and l have also challenged my closed Mentees about this.

@Theinitiatesplc have always had about 35% stake in The Initiates Uganda but the moment a disclosure was made about a further 20% stake moving it to a subsidiary like position of 55% holding, that was the first time many seems to realise. Check 2025FY doc, previous 35% stake clearly written there and not the first time it was published. NGX website is FREE and the financials of all listed companies are there. My first degree is Civil Engineering and not accounting or economics hence you can learn anything you are determined to learn. Up till today, my core expertise is #ProjectManagement predominantly but not exclusively in Engineering. Despite this, Chartered Accountants cannot tell me blue is yellow because l considerably understand their expertise, same way anyone can learn my expertise with determination. Even when my Company accounts are prepared for me by Chartered Accountants in UK and Nigeria , l still ask questions because l have a fair knowledge of what is presented to me.

If you are one of the "what can l buy investors" and keep abusing those telling you that you are threading the wrong path if indeed you are in this for your future, KONTINU !

Behind the curve investors, please read. Not just about investment but anything that adds value to you.

When many never knew TIP existed, l was already holding 11m units at an average cost of N2.06 and that was prior to the rights and public offer. Before the late comers arrived through the offers, l have already moved out about N150m profit by selling off 10.7m units while l still bought 2.2m rights at N7 per unit and 2.5m public offer at N9.50 per unit. It means l am well positioned again for what the futute holds, NOT A BUY, HOLD OR SELL ADVISE

Right now, Investors on the NGX should be positioning on very good stocks nobody is talking about instead of following noise and hype. You only make discoveries by reading and learning the proper way. Up to you how you learn. Even the Investors l mentor, l challenge them to read and can ask me questions after on what needs clarification. That is the only way to get better and stand on your own feet over time;

THE MICROWAVE GENERATION !!!

[Ben Dicken @BenjDicken]

This is why Designing Data Intensive Applications and Database Internals remain great reads.

[Codie Sanchez @Codie_Sanchez]

The biggest cheat code on the planet is internalizing that nobody cares about your life...People think about you for one second and then think about themselves again. So you might as well do what you want.

[T! @unkletayor]

@BharukaShraddha Sql

[Ronnie, Translator of Great Work🦋✨ @ronn_aa]

My unfair advantage in life is that I refuse to stay average at anything that matters to me.

If I notice a gap in my life or work, I become obsessed with closing it.

[pulkit mittal @pulkit_mittal_]

Are you stuck on Arrays and restarting 10th time now?? Do the following for DSA prep.

Find 10 EASY problems for each of these topics, on Leetcode (sorted by acceptance).

Array
Binary search
Recursion
Backtracking
Linked List
Stacks
Tree
Queues
Dynamic Programming
Graph

That makes total 100 easy problems. No overthinking, no sheets, just solve them in the order, shouldn’t take more than a month.
Then we’ll talk about your confidence.

Most important: You should’ve all problems with solutions in your notes, at one place.

50% of DSA prep is just about to keep going while feeling dumb and inadequate every single day. You stop, you loose.

[Xen @lawrenx_]

seen a lot of people not understand this.

so let’s explain.

let’s say a land in Lekki is selling for ₦10m today and you think it’ll be worth ₦20m next year but you don’t have ₦10m right now.

so you meet the landowner and say “I’ll pay you ₦900k today just to reserve my right to buy this land at ₦11m anytime within the next year.”

the owner agrees.

two things can happen:

1. land goes to ₦20m → you exercise your right, buy at ₦11m, immediately sell at ₦20m.

profit of ₦9m. all from a ₦900k deposit.

2. land stays at ₦10m or falls → your reservation expires worthless. You lose the ₦900k.

the landowner keeps it and moves on.

the person who just bought the land outright?

they spent ₦10m. Land went to ₦20m, they made ₦10m profit.

good.

but you made ₦9m with only ₦900k.

thats the power.

but if land crashes to ₦5m, they still have land worth ₦5m.

you have nothing. your ₦900k is gone.

simple rules:
- only do this when you’re very sure the price will jump
- don’t use money you can’t afford to lose
- don’t wait till the last minute to cash out

That’s LEAPS.

A reservation bet on a stock’s future price.