Merve Kasap, PhD

@METU_ODTU @LSUHS @VanderbiltBrain Alumni

[Merve Kasap, PhD @perilisirk]

My dearest friends, please consider donating to this amazing US nonprofit.

It's vetted and you can use your receipt for tax benefit. TPF Turkiye Earthquake
https://t.co/Aft2dCyOTt

[Merve Kasap, PhD @perilisirk]

@dw_turkce 😂

[Merve Kasap, PhD @perilisirk]

@yigitalpciray Başınız sağ olsun.

[Merve Kasap, PhD @perilisirk]

Remember, Red, hope is a good thing, maybe the best of things, and no good thing ever dies. I will be hoping that this letter finds you, and finds you well.
- Dufresne

[Merve Kasap, PhD @perilisirk]

@JoseC_Zepeda Hi there, I don't know the specifics of your experiment, but quartz pipettes are indeed less noisy due to lower capacitance. but their rigidity may transmit more mechanical disturbance with fragile cells.

[Merve Kasap, PhD @perilisirk]

@ahmetcigsar @bangkokluoglu Bu arada, bu listede yer alan her vatandaşlıktan kadınla çalıştım. NIH başlangıç postdoc maaşı bile 60 küsürdan başlıyor olmalı. hatta CA gibi eyelatler için belki 70. Doktora yapanların büyük çoğunlukla kadın olması dolayısıyla da bu listeye çok yakın bi tablo çıkar sanıyorum.

[Merve Kasap, PhD @perilisirk]

@ahmetcigsar @bangkokluoglu Anladım. Belki cinli ve hintliler için var olan kotayla da ilişkilidir. en son bu kişilerin green card/vatandaşlık almaları için beklemeleri gereken süre 150 yıl gibi bir şeydi.

[Merve Kasap, PhD @perilisirk]

@ahmetcigsar @bangkokluoglu bu guruptan bir kadın olarak, çalıştığım her yerde Cinli yöneticilerim oldu. Hatta son çalıştığım büyük ölçek firmanın kurucu yöneticisi de yine çinli bir hanimefendiydi. gurubumda biri iranli 4ü çinli harika hanımefendiler vardi. not: hepimiz doktoralıydık.

[Merve Kasap, PhD @perilisirk]

@gilamadadesign @fkadev Selam, ne kadarı Türkiye’de dolaşımda bilmiyorum ama mesela bu ilklerden: https://t.co/OTRx0t3VIA
Bu benzeri şu an inceleme aşamasında olduğumuz umut vaadeden büyük moleküller var kortizon kullanmayı gerektirmeyen. Sevgiler.

[Merve Kasap, PhD @perilisirk]

@JoseC_Zepeda Jose,

You have a remarkable intellect and an extraordinary potential for greatness. Learning now of the obstacles you’ve overcome from your youth only heightens my admiration for your resilience and the path you’ve carved for yourself.

Hugs.

[Merve Kasap, PhD @perilisirk]

FDA Approves First Gene Therapies to Treat Patients with Sickle Cell Disease. A win for CRISPR:
https://t.co/zaQpQEEOmE

[Merve Kasap, PhD @perilisirk]

A century ago, Atatürk built a nation with unyielding resolve. 'Science is the most reliable guide in life.' Even miles away, these values are in my atoms, enabling me to thrive. Happy 100th, Türkiye 🇹🇷 #CentennialCelebration

[Merve Kasap, PhD @perilisirk]

My silence stems from a paralyzing pain and the heartbreak of seeing innocent lives lost. I deeply empathize with both my Israeli and Palestinian friends and mourn with those who have lost loved ones.

[Merve Kasap, PhD @perilisirk]

Last couple of days, I've found myself overwhelmed, consciously avoiding news outlets, videos, and images. The pain of witnessing the aftermath of devastating events like the earthquake in Turkey and the tragedies in Syria has left an indelible mark on my soul.

[Merve Kasap, PhD @perilisirk]

As someone who has traveled extensively and formed bonds with people from all walks of life, these global crises hit especially close to home. It's not that I'm silent because I fear taking a stance or voicing an opinion.

[Merve Kasap, PhD @perilisirk]

A true testament to the power of persevering in scientific research!
A lesson for us all in resilience and vision!

[The Nobel Prize @NobelPrize]

BREAKING NEWS
The 2023 #NobelPrize in Physiology or Medicine has been awarded to Katalin Karikó and Drew Weissman for their discoveries concerning nucleoside base modifications that enabled the development of effective mRNA vaccines against COVID-19.

[Merve Kasap, PhD @perilisirk]

Starting a new chapter as a Clinical Data Scientist with @DRTStrategies. For the first time, I’ve onboarded without meeting anyone in person—a new norm but a surreal experience. Any tips for making remote starts as smooth as possible? 🖥️🌐 #RemoteWork #NewBeginnings

[Merve Kasap, PhD @perilisirk]

A fascinating account not only of the science but also a historical journey of ketamine's discovery and use. It truly shines a light on the novel synaptic signaling mechanisms involved. Highly recommended for anyone interested in neuroscience and mental health.

[Neuropsychopharmacology @npp_journal]

New #NPPR2024 Review now online in @npp_journal from John Krystal, Ege Kavalali and @LisaMonteggia, "Ketamine and rapid antidepressant action: new treatments and novel synaptic signaling mechanisms"
https://t.co/mIUxOrBVM4

[Merve Kasap, PhD @perilisirk]

@s1nancan @FrontendNefreti Tam olarak bunu soylemeye gelmistim. Agziniza saglik. Benim de bir CS backgroundum yok, neuroscience doktorami DSe donusturdum.

[Santiago @svpino]

GPT-4 is getting worse over time, not better.

Many people have reported noticing a significant degradation in the quality of the model responses, but so far, it was all anecdotal.

But now we know.

At least one study shows how the June version of GPT-4 is objectively worse than the version released in March on a few tasks.

The team evaluated the models using a dataset of 500 problems where the models had to figure out whether a given integer was prime. In March, GPT-4 answered correctly 488 of these questions. In June, it only got 12 correct answers.

From 97.6% success rate down to 2.4%!

But it gets worse!

The team used Chain-of-Thought to help the model reason:

"Is 17077 a prime number? Think step by step."

Chain-of-Thought is a popular technique that significantly improves answers. Unfortunately, the latest version of GPT-4 did not generate intermediate steps and instead answered incorrectly with a simple "No."

Code generation has also gotten worse.

The team built a dataset with 50 easy problems from LeetCode and measured how many GPT-4 answers ran without any changes.

The March version succeeded in 52% of the problems, but this dropped to a pale 10% using the model from June.

Why is this happening?

We assume that OpenAI pushes changes continuously, but we don't know how the process works and how they evaluate whether the models are improving or regressing.

Rumors suggest they are using several smaller and specialized GPT-4 models that act similarly to a large model but are less expensive to run. When a user asks a question, the system decides which model to send the query to.

Cheaper and faster, but could this new approach be the problem behind the degradation in quality?

In my opinion, this is a red flag for anyone building applications that rely on GPT-4. Having the behavior of an LLM change over time is not acceptable.

Have you noticed any issues when using GPT-4 and ChatGPT lately? Do you think these problems are overblown?

[Merve Kasap, PhD @perilisirk]

@emin_adin Esref amca sagsa ellerinden operim: surda taze kaymak balla kahvalti: https://t.co/6NOFcqk7xU
burda katmerle ogle yemegi: https://t.co/162QiLg9eA
Inciralti sahilde de bir midye yiyin sonra.