Mustafa Mert Tunalı

Building machine learning systems. AI infrastructure, distributed workloads, and LLMs.

[Mustafa Mert Tunalı @tunalim_]

@erdalgungor1905 Ben de oyle bir problem yasamadim herhangi bir Alman ile fakat Almanya'da dogup buyuyen Turk'lerde daha fazla var bu kibircilik bana kalirsa. TR'den gelenlerde ben bu dediginizi cok goremedim acikcasi. Ayristiricilik degil amacim ama TR'den gelenlerdeki kibircilik daha az.

[Mustafa Mert Tunalı @tunalim_]

@abhitwt It’s funny for now, but it won’t be funny if this happens.

[Mustafa Mert Tunalı @tunalim_]

@karpathy Congrats! How many rounds of interviews did you have to go through? 😅

[Mustafa Mert Tunalı @tunalim_]

@verstappenews I think Max pushed so hard and hit the curbs a lot, so the issue might be suspension 🥲

[Mustafa Mert Tunalı @tunalim_]

@2_sentencehrr @fatih Yurtdisi hizmet ihracati kapsaminda yararlandigim istisna tutari yuksek oldugu icin bu tasdik raporunu istiyorlar, fiyatta da buna gore degisiklik gosteriyor

[Mustafa Mert Tunalı @tunalim_]

@ardabalkandev Oyle ozen gosterdigim soylenemez ama iyi gidiyor simdilik

[Mustafa Mert Tunalı @tunalim_]

@bernhardsson @_gongy Now it’s time to visit Nurburgring 💚🌲

[Mustafa Mert Tunalı @tunalim_]

@sameenkarim @github Finally. Smaller PRs improve the quality of a code review a lot but sometimes when there are lots of PRs, it's possible to miss them among messages etc. I think this way we will be able to organize much better + automation.

[Mustafa Mert Tunalı @tunalim_]

@glennwrites1 What the

[Mustafa Mert Tunalı @tunalim_]

@paulgb lmao

[Mustafa Mert Tunalı @tunalim_]

@fejes713 That looks so smooth, congrats!

[Mustafa Mert Tunalı @tunalim_]

@f_yorumlar @AliNaciKucuk Adamlar 90+3 kirmizi kart gorup, 90+5’de gol yiyip 10 kisi ile 90+11’de gol atmis farkinda misiniz? :)

[Mustafa Mert Tunalı @tunalim_]

@p_wilfried_ @melvynx I got your point which makes sense but how is Cursor better here? I can add Claude Code as en extension to Vscode / Cursor, and use it if the case is IDEs.

[Mustafa Mert Tunalı @tunalim_]

@steipete @theo Exactly. I use different browser profiles (e.g., Chrome Personal vs. Work profile).

Glad to hear it's on the radar for the beta then!

[Mustafa Mert Tunalı @tunalim_]

CodexBar looks great! One recommendation: I use both personal and work accounts for Claude and Cursor, and I’d love to be able to see usage limits for all four separately.

Right now, it feels a bit limited because I can only track one at a time. It would be awesome if there was an "Active" toggle for each account so I don't have to manually re-configure everything in Cursor or Claude whenever I switch. What do you think? @steipete

[Mustafa Mert Tunalı @tunalim_]

@cursor_ai Experiencing frequent freezes with the new model during reasoning. It often gets stuck, and sometimes the 'details' view shows progress that isn't actually rendering in the response.

[Mustafa Mert Tunalı @tunalim_]

@DhravyaShah @supermemory Why no entity-relation-entity parsing?

[Mustafa Mert Tunalı @tunalim_]

@bosluktakaldim Istanbul'da sokakta bu sekil kedi gorurseniz Ataköy Vetra Veteriner Kliniği'ne goturebilirsiniz, ben goturdum cok sagolsunlar yardimci olmuslardi. Videoda motorcu ve kedi haric herkes saf kotuluk, umarim hak ettikleri karmayi en kisa zamanda yasarlar.

[Mustafa Mert Tunalı @tunalim_]

@fatih @cursor_ai I'm also okay to see false positives. it's always good to double check PRs after BugBot found "possible" critical issues.