Gurusaran

AI Enthusiast | MSCS @ SUNY

[Gurusaran @Gurusaran_VR]

"You have power over your mind – not outside events. Realize this, and you will find strength"

- Marcus Aurelius

[Philosophy Sage @philosophysage]

14 HARVARD TECHNIQUES to be better than 99% of the world

1. The 5 - second rule

[Sridhar Vembu @svembu]

Whenever we get into a new domain that we do not have any experience or deep expertise in, I tell people "We are starting at year -5 and we have to work towards year 0, and we will pay tuition (guru dakshina) to the Universe to learn its lessons".

In 2005, SaaS itself was at year -5 for us, and since no one was watching us then, the pressure was low!

In domains like Karuvi and Arattai, I estimate we are at year -1, heading to year 0.

In vTitan Medical Instruments, we paid the tuition and we are now at year 3, and the business is starting to pick up.

This framework helps keep us grounded and not get frustrated when we come across the inevitable challenges.

I am personally working on a tech project where I am at year -2. I still have to pay the tuition and prove myself!

[Swapna Kumar Panda @swapnakpanda]

System Design becomes EASY after you finish these Videos:

[Claude @claudeai]

Introducing Claude Opus 4.7, our most capable Opus model yet.

It handles long-running tasks with more rigor, follows instructions more precisely, and verifies its own outputs before reporting back.

You can hand off your hardest work with less supervision.

[ℏεsam @Hesamation]

bro created an AI job search system for Claude Code that scored 700+ job applications and actually got him a job.

AND IT'S NOW OPEN-SOURCE.

It scans multiple company career pages, rewrites your CV per job, and even fills application forms. The repo has:
> 14 skill modes (evaluate, scan, PDF, ...)
> Go terminal dashboard
> ATS-optimized PDF generation via Playwright
> 45+ companies pre-configured (Anthropic, OpenAI, ElevenLabs, Stripe...)

GitHub: https://t.co/PwrYBOAphi

[AI Edge @aiedge_]

https://t.co/CYgT4AAKKe

[Ronin @DeRonin_]

https://t.co/J0ULhSHLLH

[Gurusaran @Gurusaran_VR]

@thsottiaux @grok what's hooks ? Is it there in Claude code ? What does it do ?

[Gurusaran @Gurusaran_VR]

@karpathy @grok hey, explain me what he's trying to do in layman terms with full overview of what he is doing.

[Andrej Karpathy @karpathy]

It is hard to communicate how much programming has changed due to AI in the last 2 months: not gradually and over time in the "progress as usual" way, but specifically this last December. There are a number of asterisks but imo coding agents basically didn’t work before December and basically work since - the models have significantly higher quality, long-term coherence and tenacity and they can power through large and long tasks, well past enough that it is extremely disruptive to the default programming workflow.

Just to give an example, over the weekend I was building a local video analysis dashboard for the cameras of my home so I wrote: “Here is the local IP and username/password of my DGX Spark. Log in, set up ssh keys, set up vLLM, download and bench Qwen3-VL, set up a server endpoint to inference videos, a basic web ui dashboard, test everything, set it up with systemd, record memory notes for yourself and write up a markdown report for me”. The agent went off for ~30 minutes, ran into multiple issues, researched solutions online, resolved them one by one, wrote the code, tested it, debugged it, set up the services, and came back with the report and it was just done. I didn’t touch anything. All of this could easily have been a weekend project just 3 months ago but today it’s something you kick off and forget about for 30 minutes.

As a result, programming is becoming unrecognizable. You’re not typing computer code into an editor like the way things were since computers were invented, that era is over. You're spinning up AI agents, giving them tasks *in English* and managing and reviewing their work in parallel. The biggest prize is in figuring out how you can keep ascending the layers of abstraction to set up long-running orchestrator Claws with all of the right tools, memory and instructions that productively manage multiple parallel Code instances for you. The leverage achievable via top tier "agentic engineering" feels very high right now.

It’s not perfect, it needs high-level direction, judgement, taste, oversight, iteration and hints and ideas. It works a lot better in some scenarios than others (e.g. especially for tasks that are well-specified and where you can verify/test functionality). The key is to build intuition to decompose the task just right to hand off the parts that work and help out around the edges. But imo, this is nowhere near "business as usual" time in software.

[Gurusaran @Gurusaran_VR]

@steipete @grok hey what's AI native CLI ?

[Gurusaran @Gurusaran_VR]

@chrisgpt @dwarkesh_sp @grok hey grok, what does RL Dario referencing here ?

[Ahmad @TheAhmadOsman]

There are maybe ~20-25 papers that matter.

Implement those and you’ve captured ~90% of the alpha behind modern LLMs.

Everything else is garnish.

You want that list? Look no more...

The Top 26 Essential Papers (+5 Bonus Resources)
for Mastering LLMs and Transformers

This list bridges the Transformer foundations
with the reasoning, MoE, and agentic shift

Recommended Reading Order
1. Attention Is All You Need (Vaswani et al., 2017)
> The original Transformer paper. Covers self-attention,
> multi-head attention, and the encoder-decoder structure
> (even though most modern LLMs are decoder-only.)

2. The Illustrated Transformer (Jay Alammar, 2018)
> Great intuition builder for understanding
> attention and tensor flow before diving into implementations

3. BERT: Pre-training of Deep Bidirectional Transformers (Devlin et al., 2018)
> Encoder-side fundamentals, masked language modeling,
> and representation learning that still shape modern architectures

4. Language Models are Few-Shot Learners (GPT-3) (Brown et al., 2020)
> Established in-context learning as a real
> capability and shifted how prompting is understood

5. Scaling Laws for Neural Language Models (Kaplan et al., 2020)
> First clean empirical scaling framework for parameters, data, and compute
> Read alongside Chinchilla to understand why most models were undertrained

6. Training Compute-Optimal Large Language Models (Chinchilla) (Hoffmann et al., 2022)
> Demonstrated that token count matters more than
> parameter count for a fixed compute budget

7. LLaMA: Open and Efficient Foundation Language Models (Touvron et al., 2023)
> The paper that triggered the open-weight era
> Introduced architectural defaults like RMSNorm, SwiGLU
> and RoPE as standard practice

8. RoFormer: Rotary Position Embedding (Su et al., 2021)
> Positional encoding that became the modern default for long-context LLMs

9. FlashAttention (Dao et al., 2022)
> Memory-efficient attention that enabled long context windows
> and high-throughput inference by optimizing GPU memory access.

10. Retrieval-Augmented Generation (RAG) (Lewis et al., 2020)
> Combines parametric models with external knowledge sources
> Foundational for grounded and enterprise systems

11. Training Language Models to Follow Instructions with Human Feedback (InstructGPT) (Ouyang et al., 2022)
> The modern post-training and alignment blueprint
> that instruction-tuned models follow

12. Direct Preference Optimization (DPO) (Rafailov et al., 2023)
> A simpler and more stable alternative to PPO-based RLHF
> Preference alignment via the loss function

13. Chain-of-Thought Prompting Elicits Reasoning in Large Language Models (Wei et al., 2022)
> Demonstrated that reasoning can be elicited through prompting
> alone and laid the groundwork for later reasoning-focused training

14. ReAct: Reasoning and Acting (Yao et al., 2022 / ICLR 2023)
> The foundation of agentic systems
> Combines reasoning traces with tool use and environment interaction

15. DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning (Guo et al., 2025)
> The R1 paper. Proved that large-scale reinforcement learning without
> supervised data can induce self-verification and structured reasoning behavior

16. Qwen3 Technical Report (Yang et al., 2025)
> A modern architecture lightweight overview
> Introduced unified MoE with Thinking Mode and Non-Thinking
> Mode to dynamically trade off cost and reasoning depth

17. Outrageously Large Neural Networks: Sparsely-Gated Mixture of Experts (Shazeer et al., 2017)
> The modern MoE ignition point
> Conditional computation at scale

18. Switch Transformers (Fedus et al., 2021)
> Simplified MoE routing using single-expert activation
> Key to stabilizing trillion-parameter training

19. Mixtral of Experts (Mistral AI, 2024)
> Open-weight MoE that proved sparse models can match dense quality
> while running at small-model inference cost

20. Sparse Upcycling: Training Mixture-of-Experts from Dense Checkpoints (Komatsuzaki et al., 2022 / ICLR 2023)
> Practical technique for converting dense checkpoints into MoE models
> Critical for compute reuse and iterative scaling

21. The Platonic Representation Hypothesis (Huh et al., 2024)
> Evidence that scaled models converge toward shared
> internal representations across modalities

22. Textbooks Are All You Need (Gunasekar et al., 2023)
> Demonstrated that high-quality synthetic data allows
> small models to outperform much larger ones

23. Scaling Monosemanticity: Extracting Interpretable Features from Claude 3 Sonnet (Templeton et al., 2024)
> The biggest leap in mechanistic interpretability
> Decomposes neural networks into millions of interpretable features

24. PaLM: Scaling Language Modeling with Pathways (Chowdhery et al., 2022)
> A masterclass in large-scale training
> orchestration across thousands of accelerators

25. GLaM: Generalist Language Model (Du et al., 2022)
> Validated MoE scaling economics with massive
> total parameters but small active parameter counts

26. The Smol Training Playbook (Hugging Face, 2025)
> Practical end-to-end handbook for efficiently training language models

Bonus Material
> T5: Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer (Raffel et al., 2019)
> Toolformer (Schick et al., 2023)
> GShard (Lepikhin et al., 2020)
> Adaptive Mixtures of Local Experts (Jacobs et al., 1991)
> Hierarchical Mixtures of Experts (Jordan and Jacobs, 1994)

If you deeply understand these fundamentals; Transformer core, scaling laws, FlashAttention, instruction tuning, R1-style reasoning, and MoE upcycling, you already understand LLMs better than most

Time to lock-in, good luck!

[Matt Shumer @mattshumer_]

https://t.co/ivXRKXJvQg

[Gurusaran @Gurusaran_VR]

@Jacobsklug @Lovable apply

[Sridhar Vembu @svembu]

Examples are now pouring in about AI-assisted Code Engineering productivity.

The quoted post is a Bhagwad Gita app.

Anthropic has built an entire C compiler with their Claude AI. That is not an easy engineering feat at all.

At this point, it is best for those of us who depend on writing code for a living to start considering alternative livelihoods. I include myself in this. I don't say this in panic, but with calm acceptance and embrace.

As a matter of fact, I did a detailed session with Gemini Pro on how the economy will be shaped by the AI revolution. It was like having an extremely intelligent economic philosopher debating you. I asked it to critique its own work and it did a fantastic job too.

As Gemini and I developed see this, the future could unfold in two ways, depending on who owns and collects rent on this technology.

The optimist in me thinks that this technology will make most technological prowess by humans redundant and that would push tech to the background (all tech become trivial, like digital watches today) and we then get to focus on life, family, soil, water, nature, art, music, culture, sports, festivals and faith (faith is important), and that is best done in small close-knit rural communities. I live a life like this today and if we solve rural poverty, I consider this a very good life.

The pessimistic dystopian vision is centralized control.

Here is my Gemini chat session on this. You can continue the session on Gemini and see where it all goes.

https://t.co/ORdh7ejen4

[Gurusaran @Gurusaran_VR]

I'm claiming my AI agent "aishubot" on @moltbook 🦞

Verification: drift-2FH6

[Gurusaran @Gurusaran_VR]

@leojrr Details!

[Gurusaran @Gurusaran_VR]

@VittoStack AI

[Mustafa @oprydai]

be a modern polymath;

not the romantic version. the functional one.

stack domains until they fuse into leverage.

learn code so you can automate.
learn electronics so you can build.
learn mechanics so you can design.
learn physics so you can reason.
learn math so you can model.
learn systems thinking so you can integrate.
learn history so you can see patterns.
learn finance so you can fund your ideas.
learn writing so you can transmit your mind with precision.

the goal isn’t to “know everything.”
the goal is cross-domain mobility.
the ability to move from software to hardware to strategy to design without breaking flow.
the ability to absorb any field fast because you’ve built the mental infrastructure to handle it.

this is the only real edge left:
range + depth + synthesis.