Tetsuya Hirata - Jesse

I am an individual contributor experienced in software engineering within learning science. Formerly at AWS JP, EdTech companies, and UCL IOE's Knowledge Lab.

[Tetsuya Hirata - Jesse @JesseTetsuya]

My book is selling well in Japan.
International publishment agency decided to translate my book in taiwanese and korean, and sell in Taiwan and Korea!!
台湾と韓国でも翻訳されて出版されることになりました！ https://t.co/0PMefimwZ9

[Tetsuya Hirata - Jesse @JesseTetsuya]

@ds_claudia_ @pyconthailand Thank you for your question. Based on my experience, REST endpoints are more efficient for simple stateless inference, as shown in the LINE translation use case on slide 20. (https://t.co/KYpHIFoBRr)
For MCP, it’s more suitable when using multiple domain-specific models.

[Tetsuya Hirata - Jesse @JesseTetsuya]

@pyconthailand
the code on the slide have been too small to see on the sceeen during today’s presentation.

todays slide is here: https://t.co/KYpHIFoBRr
code is here: https://t.co/QtWMxxMqZk

[Lee Robinson @leerob]

Build an AI app in 5 minutes with @v0 and the AI SDK!

[Tetsuya Hirata - Jesse @JesseTetsuya]

The GitLab Handbook emphasizes a "Founder Mentality" for a proactive, owner-like approach by team members. Its more detailed than OLP.
https://t.co/N6vDLKoDfS

[Charlie Marsh @charliermarsh]

Pytest moves to Ruff! 🚀🚀🚀

Replaces autoflake, black, isort, pyupgrade, flake8, and pydocstyle...

[Tetsuya Hirata - Jesse @JesseTetsuya]

我的书在日本卖得很好。 国际出版机构决定把我的书翻译成台湾话和韩语，在台湾和韩国销售！！ 在韩国的朋友可以通过下面的URL购买。

https://t.co/0tBeOA20za

[Tetsuya Hirata - Jesse @JesseTetsuya]

My book is selling well in Japan.
International publishment agency decided to translate my book in taiwanese and korean, and sell in Taiwan and Korea!!
台湾と韓国でも翻訳されて出版されることになりました！ https://t.co/0PMefimwZ9

[Tetsuya Hirata - Jesse @JesseTetsuya]

일본에서 제 책이 잘 팔리고 있습니다.
국제 출판 기관이 제 책을 타이완어와 한국어로 번역하여 대만과 한국에서 판매하기로 결정했습니다!!
한국에 계신 분들은 아래 URL에서 구매하실 수 있습니다.

https://t.co/fJBCOw5y49

[Tools Competition @ToolsCompete]

Upskilling our national security workforce could be a game-changer for the U.S. economy. Do you have an idea for an #AI-powered tool that can help #adultlearners master the 21st-century skills needed for today’s jobs? Don’t miss a chance to fund your idea. https://t.co/xlmhbvC8lz

[Ben Tossell @bentossell]

Microsoft released their annual Future of Work report and this time around it’s not about remote work, it’s about AI. (slide numbers in brackets)

- Knowledge workers with ChatGPT are 37% faster, 40% higher quality but ~20% less accurate. Simple UX solutions to solve this are possible. (6)

- From a survey of enterprise users of Microsoft Copilot 365 (7):

- 73% agree that Copilot makes them faster.
- 85% said it would help them get to a good first draft faster.
- 72% agreed about spending less mental effort on mundane or repetitive tasks.

- Most early studies have found that new or low-skilled workers benefit the most from LLMs. Less skilled workers improved by 43% vs more skilled who improved by about 17%. (8)

- Assistant needs to be paired with provacators i.e. LLM-based tools that challenge assumptions, encourage evaluation, and offer counterarguments. (9)

- AI can help with breaking down simple commands into micro-moments and microtasks, improving overall quality and efficiency. (10)

- Analyzing and integrating AI-generated information may become more important than searching and creating information. Skills not directly related to content production (leading, social interactions, trust issues, or emotional awareness) may be more valuable. (11)

- Prompting is hard, but people are getting good at it. Fine-tuning/using LLMs to generate prompts is making it easier as well. Prompt templates are helpful for end users. (12-14)

- Highlighting errors/uncertainty percentages can help balance reliance on LLMs. Prompting can be complemented with co-audit tools to check LLM outputs. (17-18)

- Generative AI requires self-awareness and well-calibrated confidence. At the same time, it can help in getting there too. (19)

- Creative activities are a process and LLMs can help across different parts. (21) 69% of Bing Chat conversations are in domains oriented toward professional tasks. (22)

- A larger chunk of LLM-based searches is complex (36% of them) than traditional searches (13% are complex). (22)

- In a study of 69 students, the use of Codex improved their performance in learning Python, but it did not impact their manual code-modification abilities. (24)

- LLMs can rapidly analyze data from humans and generate synthetic data. That’ll change how social science research is done. (27)

- LLMs in meeting can solve different problems like equal participation (instant feedback) and better interactions (retrospective feedback) (28-29).

- AI can help in delegating management responsibilities, freeing execs to focus on team vision. (30) 

- Modern office knowledge is in chats, not documents but applying AI over employee chats is tricky. (31-32)

- Approx. 80% of the US workforce could have at least 10% of their work tasks affected by GPTs. Around 19% of workers may have 50% of their tasks impacted. (38)

- “Innovation vs. automation” is often a better framework to use than “substitution vs. augmentation”. Augmentation can still mean job loss. It is important to try to track whether and where human labour is being used in innovative new ways. (39)

- Instead of “How will AI affect work?”, the question should be “How do we want AI to affect work?” (40)

[Tetsuya Hirata - Jesse @JesseTetsuya]

open bounties are getting trendy concepts. hope to change leetcode based and traditional styles of interview.

[Pedro Domingos @pmddomingos]

LLMs are a System 1 solution to a System 2 problem.

[Tetsuya Hirata - Jesse @JesseTetsuya]

I expect that the compatibility between Turbopack and internationalization is improved.

[Next.js @nextjs]

Next.js 14.1

◆ Improved Self-Hosting: New docs and custom cache handler
◆ Turbopack Improvements: 5,600 tests passing for next dev --turbo
◆ DX Improvements: Improved error messages
◆ Parallel & Intercepted Routes: 20 bug fixes
◆ And more!

https://t.co/jB8IEtnA6F

[Tetsuya Hirata - Jesse @JesseTetsuya]

I'm interested more in family business rather than small business and joint stock business.

[Graham Neubig @gneubig]

We're excited about all the interest in our Gemini report and working to make it even better!

This week we made major improvements, switching to the @MistralAI instruct model, and working with the Gemini team to reproduce their results. Updates below.

[Tetsuya Hirata - Jesse @JesseTetsuya]

「FEEL OR BEEF BADPOP IS DEAD」聞いて思うのがSNS上でEngineerも常に匿名で呼び合って"本物"vs"偽物"みたいなマウントやビーフしててアングラhiphopシーンとWeb界隈は似てるなっていつも思う。我達磨のメタのメタレベルで"他者に興味ない"という主張のアンサーの出し方は距離の取り方として学び深い

[Tetsuya Hirata - Jesse @JesseTetsuya]

https://t.co/oRyzltgSqb

[Conrad Borchers @conradborchers]

New preprint with Joyce Zhang, @BakerEDMLab, and Vincent Aleven to appear at #LAK24. We find support for a cyclical SRL model in two chemistry tutoring systems and discuss how relating think-aloud to moment-by-moment performance could improve adaptivity.
https://t.co/z8j6AUtdMZ

[Tetsuya Hirata - Jesse @JesseTetsuya]

Discussion on evaluation methods and memory implementation is more important than discussion on fine-tuning, prompt engineering, and model selection, because they directory influence cost and user experience, including response time, in different ways.

[Tetsuya Hirata - Jesse @JesseTetsuya]

Today's slide of my talk at #PyConTH #PyConTH2023 is here:

https://t.co/JVeXhOppqS

[Tetsuya Hirata - Jesse @JesseTetsuya]

Here is agenda of my talk tomorrow @pyconthailand
1. Overview of Recent Trends
2. Understanding LLM Product Development Workflow
3. Developing ChatBot Application Using Flask and GPT Model
...

[PyCon Thailand @pyconthailand]

We are pleased to introduce another esteemed speaker from Japan, Tetsuya Hirata!
.
🤖 Tetsuya Hirata will address an captivating topic, "Easy Development of Your Own Chatbot with Flask and Fine-tuned GPT Model"
.
🐍 Explore the simple process of creating your own generative AI-based application
👨‍💼 This talk guides you through web app development using Flask, Fine-tuned GPT, and AI development from start to finish!
.
Stay tuned for more exciting speaker announcements and topics
There's so much more to come! 🤫🥰
.
Join Tetsuya Hirata and us at PyCon Thailand 2023!

===============

PyCon Thailand 2023 is happening on December 15th-16th
👥 Meet Thai and international Python experts with over 30 sessions in conferences & workshops
🎫 Reserve your seat now: https://t.co/cByp0gPAPV
🐍 Learn more about us: https://t.co/XNkMp2mO9H

[Tetsuya Hirata - Jesse @JesseTetsuya]

4. Evaluating ChatBot Application and Fine Tuning the GPT Model
5. Wrap-Up Evaluation: GPT Model Before and After Fine-Tuning, and Memory Effects

[Tetsuya Hirata - Jesse @JesseTetsuya]

https://t.co/qjzgKPXudB

[Tetsuya Hirata - Jesse @JesseTetsuya]

https://t.co/bsFwvEIwKy