Ashik Shaffi

[Unreal Engine @UnrealEngine]

Introducing Epic’s version control system: Lore! Built from scratch and open-sourced, we made scalability and performance core tenets from day one.

We’re releasing Lore as open source today.All source code, along with documentation of data formats and protocols are released under an MIT license and free to use. Help shape it into something even more powerful, together: https://t.co/zplVMjkRof

[Ashik Shaffi @AshikHafi]

Feels like the task state after compaction is preserving "continue auditing" but not converging toward "produce report".

@OpenAIDevs

[Ashik Shaffi @AshikHafi]

GPT 5.3 Codex Spark got stuck in a compaction loop

After context compaction, it seems to be repeatedly entering a loop:

"I'm doing one more audit pass"
"I'm checking a few more files"
"I'll summarize findings shortly"

@thsottiaux @jxnlco

[Ashik Shaffi @AshikHafi]

Then context compacts again and it restarts the same audit plan.

Result:

~36 minutes of runtime
dozens of file reads/searches
multiple compactions
never produced the actual findings

[Ashik Shaffi @AshikHafi]

@itisprashanth Stick to films ffs

[Ashik Shaffi @AshikHafi]

fuck! claude code is soo back.

[Arsenal @Arsenal]

WHAT A START!

Havertz races through and smashes one into the roof of the net 💥

🔵 0-1 🔴 (6)

[Arsenal @Arsenal]

It all comes down to this.

[Ashik Shaffi @AshikHafi]

@mrdbourke @ArtificialAnlys of course he does haha! Good to see you here.

[Raf 🇧🇩 @rafsenal]

Shoutout to the brother who started selling mango lassi outside the ground at 12am. Grind never stops.

[Ashik Shaffi @AshikHafi]

I fucking love this club

[Arsenal @Arsenal]

This belongs to all of us.

[Arsenal @Arsenal]

Arsenal: 2025/26 Premier League champions.

[Ashik Shaffi @AshikHafi]

@shardul_mane Should’ve hit me up!

[wh @nrehiew_]

Frontier LLMs are doing too much when it comes to editing code.

I'm excited to share this work on the Over-Editing problem which refers to models modifying code beyond what is asked of them.

The main findings are:
- Many frontier models Over-Edit with GPT 5.4 being the biggest culprit
- Reasoning models have a higher natural tendency to Over-Edit compared to their non-reasoning counterparts
- RL is the best approach to train models to perform minimal code editing while preventing catastrophic forgetting compared to SFT, DPO and Rejection Sampling.

Blog and details below!

[Ashik Shaffi @AshikHafi]

@JcCastaneda05 @theo @thsottiaux Noticed for me as well

[Mind of Saka @WFAFOOTBALL14]

“Pressure is a Privilege. And if you’re feeling any pressure or weight of expectation, you are breathing rare air, that very few of us get to live inside”.

Unbelievable piece of poetry .

[SG @04gully]

Oi why he lowkey putting the pressure on us, we not built for that shit bro

[Ashik Shaffi @AshikHafi]

@vboykis thanks for the reminder.

[dfkm @domcomps]

Not too late for skelly to buy an e scooter and go phone snatching in central london

[Arsenal @Arsenal]

Our FA Cup journey ends here.

Attentions turn to Sporting CP on Tuesday.