Ismail Cherri (@IsmailCherri) profile picture

Frontend Engineer @SonarCloud. ML/DeepL enthusiast. Searching for the next big thing.

Germany
Joined March 2011
Photo by Ismail Cherri (@IsmailCherri)Photo by Ismail Cherri (@IsmailCherri)Photo by Ismail Cherri (@IsmailCherri)Photo by Ismail Cherri (@IsmailCherri)Photo by Ismail Cherri (@IsmailCherri)Photo by Ismail Cherri (@IsmailCherri)Photo by Ismail Cherri (@IsmailCherri)Photo by Ismail Cherri (@IsmailCherri)Photo by Ismail Cherri (@IsmailCherri)Photo by Ismail Cherri (@IsmailCherri)
لو بتستخدم الIPTV وزهقت من البرامج الموجودة زي IPTV Smarters وكدة.. أنا كمان زيكم وعشان كدة عملت YAIP
18
7
208
41,381
After working intensively with both #Codex Sol 6.1 and 6.2, I can confirm with 100% certainty that there is a HUGE design flaw in its architecture, training or reasoning in terms of coding capabilities compared to #Claude! 🧵
1
1
2
689
In addition, it looses focus WAY too easy, and once it reaches this state, there is no way to get it back on track.
1
44
@OpenAI you really need to work on these issues if you want to stay relevant in the coding space. Disclaimer: I wasn't paid by anyone for this opinion, I'm just another unsatisfied customer.
1
34
Ismail Cherri retweeted
Here’s the full walkthrough: inbox → analysis → guided diff → the new PR map → Submit Review. Here’s everything that shipped since 0.4 🧵
PR review shifted: with agents writing the code, the job is checking intent (does it do what the description says?), not reading every line. So I built BetterLGTM • Checks each PR claim against the code • Guided review for huge diffs Free, open source, local.
2
1
6
302
Ismail Cherri retweeted
Can confidence keep up with AI-generated code? That's the question behind our 2027 State of Code Developer Survey, out now. The results point to verification and governance, critical as agentic software factories are set to grow 7x next year. Details: ow.ly/8xV8510647A
1
1
5
312
Ismail Cherri retweeted
PR review shifted: with agents writing the code, the job is checking intent (does it do what the description says?), not reading every line. So I built BetterLGTM • Checks each PR claim against the code • Guided review for huge diffs Free, open source, local.
10
16
5
127
11,315
الدنيا مش مستاهلة الواحد يعمل بوست على تويتر
1
1
137
Sure it does.. Cursor sold their soul when they sold out to the Tusk
Grok bot sucks!!!
1
122