Alexey Iskrov

Born in 🇺🇿
Went to school in 🇺🇦
Graduated from university in 🇷🇺
Immigrated from 🇰🇿 to 🇨🇦

[Alexey Iskrov @AlexeyIskrov]

Been to the 80% of the West Coast. Totally agree with this statement. BTW it continues in Costa Rica. Never been to Nicaragua and Panama, but I'd include them as well.

[James🗳 @_fat_ugly_rat_]

This is genuinely the best part of our entire planet

[Alexey Iskrov @AlexeyIskrov]

@davidasinclair Wow, people have no idea how budgets are allocated in Russia. Just google how much Russia spent on "nano technologies" and try to find any science breakthrough in this space coming from Russia. This is just a money grab from somebody at the top. Pure corruption.

[Alexey Iskrov @AlexeyIskrov]

@sandislonjsak Addictive productivity

[Alexey Iskrov @AlexeyIskrov]

I call it addictive productivity. Never worked more in my life.

[Sandi Slonjšak @sandislonjsak]

Haven’t met anyone working less after adopting the AI.

[Alexey Iskrov @AlexeyIskrov]

So cool to build right now. You are limited just by your imagination.

[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.

[Alexey Iskrov @AlexeyIskrov]

@ylecun This is unimaginable that there are people who can support this. Poor evil.

[Yaroslav Bulatov @yaroslavvb]

Today I spent too long to find this picture I took in Japan years ago, since Google Photos doesn't let you search for monkeys! The reason is that in 2015, it classified a black dude as a gorilla. I remember the escalation, it was a sister team and they fixed it by removing simians from the label set. 2015 is ancient history and image recognition has improved in the last 11 years. It's time to #bringbackthemonkeys @JeffDean

[Alexey Iskrov @AlexeyIskrov]

@Vivers51 I’m so sorry for your loss. I don’t know what you’re going through, but I imagine your husband and son would want you to keep living, to enjoy life when you can, and to find comfort in the little things that make you happy.

[@levelsio @levelsio]

Creator of @nodejs says humans writing code is over

[Alexey Iskrov @AlexeyIskrov]

@Jason Guess what wasn't on Canada's cards?

[Alexey Iskrov @AlexeyIskrov]

@shiri_shh 👋

[Alexey Iskrov @AlexeyIskrov]

@NikkiHaley The only country that put North America at serious risk is the US.

[Alexey Iskrov @AlexeyIskrov]

@msvivianrabbit February 27, 2014

[Alexey Iskrov @AlexeyIskrov]

Absolute truth. They don't care about future of the world, they all going to die in the next 10-15 years. People over 70 shouldn't be allowed to lead any country. https://t.co/Keg0s8m9cI

[Mohit Chauhan @mohitlaws]

Putin: 73 years old
Trump: 79 years old
Xi Jinping: 72 years old
Khamenei: 86 years old
Netanyahu: 76 years old
Narendra Modi: 75 years old

All these leaders have lived their lives and are now destroying world peace and the future of coming generations.

The world needs young leaders.

[Alexey Iskrov @AlexeyIskrov]

@KatKanada_TM That thick Russian accent though. If you don't like Canada just go back to where you came from or move to US.

[Andrej Karpathy @karpathy]

I am unreasonably excited about self-driving. It will be the first technology in many decades to visibly terraform outdoor physical spaces and way of life. Less parked cars. Less parking lots. Much greater safety for people in and out of cars. Less noise pollution. More space reclaimed for humans. Human brain cycles and attention capital freed up from “lane following” to other pursuits. Cheaper, faster, programmable delivery of physical items and goods. It won’t happen overnight but there will be the era before and the era after.

[Alexey Iskrov @AlexeyIskrov]

Tried to register a new Apple dev account. Was blocked because they wouldn't accept my address. Basically they didn't accept "British Columbia" as my province from THEIR OWN drop-down menu. How do all Apple users accept this kind of service? Locked out for wrong address? Really?

[Alexey Iskrov @AlexeyIskrov]

The Zen of Python Objects 🐍

Everything is an object.
- object gives existence
- type gives form
- self is the mirror
- __dict__ is memory
- __init__ is birth
- __del__ is silence
Between them — Python lives.

[Robert Youssef @rryssf]

Holy shit... Tencent researchers just killed fine-tuning AND reinforcement learning in one shot 😳

They call it Training-Free GRPO (Group Relative Policy Optimization).

Instead of updating weights, the model literally learns from 'its own experiences' like an evolving memory that refines how it thinks without ever touching parameters.

Here’s what’s wild:

- No fine-tuning. No gradients.
- Uses only 100 examples.
- Outperforms $10,000+ RL setups.
- Total cost? $18.

It introspects its own rollouts, extracts what worked, and stores that as “semantic advantage” a natural language form of reinforcement.

LLMs are basically teaching themselves 'how' to think, not just 'what' to output.

This could make traditional RL and fine-tuning obsolete.

We’re entering the “training-free” era of AI optimization.

[Millie Marconi @MillieMarconnni]

Meta just did the unthinkable.

They figured out how to train AI agents without rewards, human demos, or supervision and it actually works better than both.

It’s called 'Early Experience', and it quietly kills the two biggest pain points in agent training:

→ Human demonstrations that don’t scale
→ Reinforcement learning that’s expensive and unstable

Instead of copying experts or chasing reward signals, agents now:

- Take their own actions
- Observe what happens
- Learn directly from consequences — *no external rewards needed*

The numbers are wild:

✅ +18.4% on web navigation (WebShop)
✅ +15.0% on complex planning (TravelPlanner)
✅ +13.3% on scientific reasoning (ScienceWorld)
✅ Works across **8 environments**

And when you add RL afterward?

🔥 +6.4% better than traditional pipelines.

Two key ideas make it work:

1. Implicit World Modeling - agents predict what happens next, forming an internal world model.

2. Self-Reflection - they compare mistakes to experts and explain why the expert choice was better.

Both scale. Both are reward-free.

Efficiency is absurd:

1/8 of expert data
86.9% lower cost
Works from 3B → 70B models

This isn’t incremental.

It’s the bridge between imitation learning and true autonomous experience.

AI agents can now teach themselves - no human hand-holding required.