Igor Berman

[Igor Berman @IgorBerman]

@lifeof_jer @cipherstein you might want to introduce token rotation. 1 year old token seems like security risk, besides what happened

[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.

[Java.IL @java_il]

Learn how to turn your load balancing to be cache-aware to shave off milliseconds in a very high rate from @taboola : https://t.co/bfvwyYBueA

[Yann LeCun @ylecun]

This is huge: Llama-v2 is open source, with a license that authorizes commercial use!

This is going to change the landscape of the LLM market.
Llama-v2 is available on Microsoft Azure and will be available on AWS, Hugging Face and other providers

Pretrained and fine-tuned models are available with 7B, 13B and 70B parameters.

Llama-2 website: https://t.co/PKrrXgHdem
Llama-2 paper: https://t.co/aINNrXNhMb

A number of personalities from industry and academia have endorsed our open source approach: https://t.co/N7HwgW9Suh

[AppsFlyer Engineering @AppsFlyerDev]

Join us for an exclusive meetup we're joining forces with @taboola on, where @IgorBerman and @HarelOpler will share their insights on scaling and optimizing data infrastructure. More info and registration here: https://t.co/9D8TW3ke1B

[Igor Berman @IgorBerman]

https://t.co/Rzqzx6capf

[Jack Vanlightly @vanlightly]

Redpanda bring out benchmark after benchmark claiming performance superiority over Apache Kafka. I decided to run my own tests to see if any of it was true.
https://t.co/hb8euOGYmB

[Igor Berman @IgorBerman]

https://t.co/rRjWYSYleQ must see for every java dev

[Daniel Kelley @danielmakelley]

Here are 27 ways to learn ethical hacking for free:

1. Root Me — Challenges.
2. Stök's YouTube — Videos.
3. Hacker101 Videos — Videos.
4. InsiderPhD YouTube — Videos.
5. EchoCTF — Interactive Learning.

[Igor Berman @IgorBerman]

https://t.co/Y0LjWcpKsm
this is most annoying bug(in our opinion) that we experienced with #ApacheSpark until now

[Igor Berman @IgorBerman]

Surviving Spark Upgrade in Production https://t.co/JoDc4cpEQh

[Apache Spark @ApacheSpark]

[ANNOUNCEMENT] Congrats to the Apache Spark community and all the contributors! The Apache Spark 3.3 is here. Try it out! https://t.co/OmAKmxv0l6

[Igor Berman @IgorBerman]

@tizkiko יש לי אחת שכבר במייל עשירי לפחות. הכל הולך לפח ישר...אומנם בנתיים לא יצרתי פילטר אוטומטי...

[Igor Berman @IgorBerman]

https://t.co/Rwg7N1M17B
relatively recent work in Taboola.
Sometimes to do 100 full table scans is not an option

[Igor Berman @IgorBerman]

@tizkiko only in java?

[Igor Berman @IgorBerman]

https://t.co/2qZ1kfGBBI

[Igor Berman @IgorBerman]

Summary of June 8 outage https://t.co/z2a8rZ6IKf
My question is: why bad configuration of one customer put #FastlyDown
IMO, bugs happened, happening and will happen and there is no practical way to eliminate them completely(it's just not cost effective)

[Igor Berman @IgorBerman]

https://t.co/EJOKvQS8By

[Igor Berman @IgorBerman]

@tizkiko @shakedko @kubernetesio my most frequent is kgp from here https://t.co/Wt7SfxSWyB

[Igor Berman @IgorBerman]

@jjirsa @liorchaga @tizkiko @taboola @cassandra I think the missing detail here is that we manage 'retention' sort of manually by keeping daily tables and in some case - hourly table. And then dropping full table when it out of retention window