Sami Ahma-aho

Digital Lead in Data Science & AI for Neste. Data Science is my business. Photography and Science is my leisure. Views are my own.

[Rowan Cheung @rowancheung]

Google just revealed Gemini and will directly integrate the AI into Google apps.

The GPT-4 competitor comes in 3 models — Ultra, Pro, and Nano.

Here's a thread of EVERYTHING you need to know:

[Peter Wang 🦋 @pwang]

Um... I just had like a 20 minute conversation with ChatGPT about the history of modern physics. If I had this shit as a tutor during high school and college.... OMG.

I think we can basically re-invent the concept of education at scale. College as we know it will cease to exist.

[Sami Ahma-aho @Samjum]

So noticed this is on the way. Booked.
#GoogleCloudNext ’21 is coming, October 12-14, 2021. https://t.co/MOzEIBSJbc

[Puolustusvoimat @Puolustusvoimat]

Taisteluämpäri M21 on ensimmäinen varuste-esine, joka jaetaan ilmaiseksi kaikille reserviläisille.
#puolustusvoimat #intti #tstämpäri

[Sami Ahma-aho @Samjum]

And to be clear: absolutely developing different solutions, benchmarking and comparisons are useful. And they come across way better when there is a genuine goal to push the technological envelope together.

[Sami Ahma-aho @Samjum]

Be inactivish on Twitter for 4 years. Come back. Notice that there is still this "we got data.table beat" movement up and kicking. BUT even more weird is the amount of students I keep running to that haven't heard about basic packages like #rdatatable #rcaret #rdrama

[Dirk Eddelbuettel @eddelbuettel]

@MattDowle Confirming the finding: data.table::fread still wins.

#rstats #rdatatable

[Sami Ahma-aho @Samjum]

My theme of the month right here. Might turn out to be my theme for the year. I venture one step further and say that neither data or technology or algorithms are the biggest obstacles for most organisations in applying data & analytics successfully.

Thi…https://t.co/zpAior7vJ2

[lak lakshmanan @lak_luster]

This is game changing in terms of simplicity. Apache Beam is powerful, but the syntax could be off-putting. But now you can write Beam pipelines with the Pandas API! Also greatly simplifies productionizing ML workflows. Experiment in Pandas, run on Dataflow

[Chip Huyen @chipro]

ML community: * create algorithms that optimize for a single objective *

Companies: * use ML to optimize for user engagement, which learns to favor extreme content since it gets the most attention *

ML community: “Why are people on Twitter getting so extreme?”

[Sami Ahma-aho @Samjum]

My #Fallout themed console at the cabin -project just got refactored due to new hardware requirements: https://t.co/j6g6k3xqVp I hope the "Product Owner" understands the "dev teams" change of heart.

[Sami Ahma-aho @Samjum]

@KatriTaskinen Kun ääntenlaskennassa kestää niin onko tässä olemassa riski että joudut paistamaan niitä vielä viikonlopun yli ;)

[Sami Ahma-aho @Samjum]

Have to read this. Expectations are high after spending considerable amount of time in the past around this. And must admit: the practicalities of the problem always superseded the tech approach

[H2O.ai @h2oai]

Download our white paper “The Benefits of Budget Allocation with AI-Driven Marketing Mix Models” to learn more about how the latest in #AI technologies reinvented #media and #marketing analytics: https://t.co/n2a7ZRrZ16 #EnterpriseAI #MachineLearning #AITransformation

[Sami Ahma-aho @Samjum]

Excited about tomorrow being able to talk with some bright minds at the @MKollektiivi organised #Martech & #Data event. The topic will be challenging: "How to succeed in Martech & Data procurement" #markkinointikollektiivi

[Chip Huyen @chipro]

2018: BERT looks good on leaderboards but it’s not practical because it's too big too slow

2020: BERT is used in almost every English search on Google

There are many engineering challenges to bring large ML models to production, but they are being solved at a great speed 🚀

[Sami Ahma-aho @Samjum]

@lak_gcp I think it just adds mystery or emphasis :) When this happens I tend to think about the SO blank about me: "Apparently, this user prefers to keep an air of mystery about them".

[Sami Ahma-aho @Samjum]

@chipro Conclude early but continue to help if useful and not patronizing. If you can, help the person realize where they are lacking skills and thinking. This way you both still get “a win” out from the discussion/interview.

[Sami Ahma-aho @Samjum]

Just saw the Clapboard Jungle by @UnstableGround at @nightvisionsff. Fantastic description of struggles moving towards your vision -one brick at a time. Thank you.

[Jordan Tigani @jrdntgn]

This morning, we launched the Public Datasets for COVID-19 program, making COVID-related data available and free to query via BigQuery (no credit card needed with the BQ sandbox). Details (and fine-print) here: https://t.co/FaKI0Y9y84

[Sami Ahma-aho @Samjum]

Birthday is on the 25th so I don’t mind if you happen to hit with news somewhere around that 😁

[Tino Tereshko 🇺🇦 @thetinot]

This is shaping up to be a huge month for BigQuery customers and the team!

[Andrej Karpathy @karpathy]

My smart coffee mug warmer’s companion app demanded I make an account, and now wants to send me push notifications and have access to my location. Amazing.