Talha Oz

Applied Scientist @Amazon
PhD from @CSS_GMU
Founder of @Twlets
Alumnus of @bogazici_cmpe & @izmirfenlise

[Türkiye Yazma Eserler Kurumu (TÜYEK) Başkanlığı @yekgovtr]

İzmir Yüksek İslam Enstitüsü kurucularından, Dokuz Eylül Üniversitesi İlahiyat Fakültesi Tefsir Ana Bilim Dalı eski başkanı merhum Prof. Dr. Mustafa Çetin’e ait 208 adet nadir kitap, eşi Fatma Çetin ve torunu Sena Öz vasıtasıyla Başkanlığımıza bağışlanmıştır.

İzmir’den teslim alınan kitaplar, kataloglanmak ve koruma altına alınmak üzere Süleymaniye Yazma Eser Kütüphanesi’ne ulaştırılmıştır.
Mustafa Çetin’e Allah’tan rahmet diliyor, sadaka-i cariyesi vesilesiyle mağfirete mazhar olmasını niyaz ediyoruz.

[Talha Oz @tozCSS]

@OneManSaas @huang_chao4969 I believe that number comes from DoL/ONet* and OpenAI’s GDPVal, which are written by professionals working in those fields. You need to read the gdpval paper.

[Talha Oz @tozCSS]

@futuristufuk Micro derken, yazar operationalized “being social” in terms of interactional signatures: reciprocity and repeated mutual exchange, sustained conversational subthreads, and conversational persistence beyond one-off replies.

[Talha Oz @tozCSS]

@futuristufuk Güzel açıklama ve sonuç — Bu YZler sosyal simülasyon için kullanılıyordu ama böyle bir field experiment yoktu. Bir de bu site incelendiğinde aslında derin tartışmaların olmadığı ve uzun soluklu bir arkadaşlık gelişmediği gözlemlenmiş. Makrodan çok mikro analiz ilginç:

[David Holtz @daveholtz]

paper (as i said, very much a draft, will update): https://t.co/QANqrBj4p7
replication code: https://t.co/fQ1yzXPeKl

additional caveat: both the coding and writing for this project were done with heavy assistance from an ai agent, which felt appropriate

[Talha Oz @tozCSS]

@okcknet @utkusen Extension’lar senin özgürlük alanını daraltıyor. Mesela VS Code bazı API’ları açmıyor extension’lara. Eğer senin ürününün başarısı için gerekli bir özelliği desteklemiyorsa, mecburen daha temelden ürünü inşa etmen gerekiyor.

[Talha Oz @tozCSS]

Paper: https://t.co/9xjk4OtwnP

[Talha Oz @tozCSS]

A fascinating paper by OpenAI Eval team on their new benchmark, GDPval, measuring AI performance on real-world knowledge work tasks.
Here are the most interesting highlights for me:

* Blinded expert-eval. Top model's (Claude) win-rate over humans: 47.6%. It means …

[Talha Oz @tozCSS]

* 1320 super realistic real-world tasks (44 occupations x 30 tasks) in their private dataset, they released 220 (44x5) tasks gold dataset on Huggingface. Average task takes 7 hours and costs $391 in professional time. These occupations represent $3T in annual economic activity.

[Talha Oz @tozCSS]

33% of scientific findings would be incorrect had researchers used LLMs instead of humans for labeling: says https://t.co/g2gf0XrOz7

This paper sparked an idea: creation of SocSciBench, a leaderboard, ranking LLMs’ annotation quality for social science hypotheses testing.

WDYT?

[Talha Oz @tozCSS]

@amir_mahla Great work. Would you mind sharing the cost for each training phase in terms of gpu and time? Wondering the resources needed to run the recipe? Ty!

[Talha Oz @tozCSS]

@SYOO86 @deedydas @SumukhV definitely a "drop-in remote worker" is on the way. An agent that can message over Slack, join Zoom meetings, read company docs, use software, makes pull requests, give daily updates, and independently complete the project. https://t.co/DxnV7MstZM

[AI at Meta @AIatMeta]

🎥 Today we’re premiering Meta Movie Gen: the most advanced media foundation models to-date.

Developed by AI research teams at Meta, Movie Gen delivers state-of-the-art results across a range of capabilities. We’re excited for the potential of this line of research to usher in entirely new possibilities for casual creators and creative professionals alike.

More details and examples of what Movie Gen can do ➡️ https://t.co/M19x2ndwnr

🛠️ Movie Gen models and capabilities
Movie Gen Video: 30B parameter transformer model that can generate high-quality and high-definition images and videos from a single text prompt.

Movie Gen Audio: A 13B parameter transformer model that can take a video input along with optional text prompts for controllability to generate high-fidelity audio synced to the video. It can generate ambient sound, instrumental background music and foley sound — delivering state-of-the-art results in audio quality, video-to-audio alignment and text-to-audio alignment.

Precise video editing: Using a generated or existing video and accompanying text instructions as an input it can perform localized edits such as adding, removing or replacing elements — or global changes like background or style changes.

Personalized videos: Using an image of a person and a text prompt, the model can generate a video with state-of-the-art results on character preservation and natural movement in video.

We’re continuing to work closely with creative professionals from across the field to integrate their feedback as we work towards a potential release. We look forward to sharing more on this work and the creative possibilities it will enable in the future.

[Talha Oz @tozCSS]

@skyzone we have been trying to reach you via "contact us", email, and phone in the past three months, nobody is responding. Please cancel our elite membership and refund us the last two months.

[Talha Oz @tozCSS]

@stikves Yeah. I am working on a related project at AWS.

[Talha Oz @tozCSS]

When assigned a task on JIRA, an AI Worker will soon be able to finish the task assigned to it just as a human employee would do. It will create a roadmap, read the company docs, implement code, send PRs, communicate with others over Slack, join Zoom +

[Talha Oz @tozCSS]

A great read on the topic is https://t.co/adXsLGU3PJ

[Talha Oz @tozCSS]

I believe by the end of 2026, these remote AI workers will be much more capable, and then it would make more sense for some companies to hire them instead of humans for many roles. What do you think?