@garrytan This has been my experience as well. Plan with Opus, execute and code review with GPT-5.4.
English
Samuel W. Failor
51 posts
Samuel W. Failor
@sfailor
Data Scientist and AI/ML Engineer - focused on AI for internet safety
United Kingdom Katılım Şubat 2009
509 Takip Edilen132 Takipçiler
I added a /codex skill that does plan and code review in Claude Code and it has found so many more bugs and helped with far more elegant architectural designs
And tomorrow I will launch native Codex support so you can use GStack on Codex
Garry Tan@garrytan
OK Codex is GOAT at finding bugs and finding plan errors
English
Samuel W. Failor retweetledi
Over a year ago, I made several criticisms of a viral article that claimed that microplastics had now accumulated to roughly 7g of our brains. For example I pointed out they gave very little detail on how they checked for contamination.
How has the study held up?
Not very well!
Saloni@salonium
I read this recent 'microplastics in the brain' article and I have a lot of questions. 🧵
English
Samuel W. Failor retweetledi
1 Start the worst European war since 1945 and commit every imaginable war crime
2 Decide how you want to be rewarded for your aggression
3 Leak the list of your desiderata to the US press
4 Wait for the US president to enforce your wish list on your victim
Russian policy 2025
English
Samuel W. Failor retweetledi
🔬 PhD opportunity alert! Join us at the @CropSciCentre @camplantsci to decode sensor–helper NLR communication mechanisms in plant immunity. Apply by 14th November! 🔍 More info 👉 findaphd.com/phds/project/d… #PhD #PlantScience #Biology Please RT
English
Samuel W. Failor retweetledi
Samuel W. Failor retweetledi
Samuel W. Failor retweetledi
MAGA is making the claim that if you watch the entire 50 minute video, Zelensky is the antagonist.
This isn't true. At all. I was with Zelensky before this meeting and it's important to know his mindset heading in.
1/ A short 🧵on how the entire ambush went down.
Jordan Schachtel@JordanSchachtel
If you watch the full Trump-Zelensky press conference, it is very clear that Zelensky, not Trump or Vance, became the antagonist. Both POTUS and VP were very respectful and cordial until Zelensky very publicly ignited a firestorm. It all starts at 40:30 1) Zelensky essentially rejects how VP described the mandate of POTUS to conduct foreign affairs, and he insinuates that Trump term one did nothing to stop Putin. 2) He then basically tells Vance that his ideas are faulty and that the administration's diplomacy won't work. These two comments are *deliberately antagonistic.* Everything was all well and good, but Zelensky took two major shots in a public forum, and they had to respond. And respond they did. Recall, this is the guy who interfered in our electoral politics and called VP "too radical," and bashed Trump in an interview with New York Mag weeks before the election. Zelensky is ENTIRELY at fault here. 100%.
English
Samuel W. Failor retweetledi
Believe in Ukraine. Stand with Ukraine. Glory to Ukraine!
English
Samuel W. Failor retweetledi
New 20hr bootcamp on Probability & Statistics!!!
Videos released weekly but full playlist already posted:
youtube.com/watch?v=sQqnia…
Probability & Statistics are cornerstones of data science and machine learning. This course rapidly covers the basics and gets into advanced topics.
YouTube
English
Samuel W. Failor retweetledi
Spike sorting with #Kilosort in #python w/ #pytorch + @pyqtgraph !
>>> pip install kilosort[gui]
To learn more, check out the docs (kilosort.readthedocs.io/en/latest/ by Jacob Pennington), or watch the talk (youtube.com/watch?v=LTSmoA… by @marius10p
)
YouTube
Marius Pachitariu@marius10p
The Kilosort4 paper is now published @naturemethods. We put a lot of work into building and properly documenting all things Kilosort (including for older versions). w/ @computingnature #kilosort paper: nature.com/articles/s4159… code: github.com/MouseLand/Kilo…
English
Samuel W. Failor retweetledi
🚨Awesome news alert🚨
Our Computational Neuroscience & Machine Learning Course is back!!
When: 8-24th July 2024
Where: CMU-Africa campus, Kigali, Rwanda
Applications open: 15th January 2024
Applications close: 15th February 2024
For news on the course follow @trend_camina
English
Samuel W. Failor retweetledi
🌱 Exciting Opportunity Alert! 🌱
We're searching for a passionate #ResearchAssistant to join our dynamic team! 🧪
Find out more below ⬇️ Only 6 days left to apply!!
jobs.cam.ac.uk/job/43348/
@CropSciCentre @plantsci
English
Samuel W. Failor retweetledi
I am delighted to advertise a 2-year postdoctoral research associate position in my lab @CropSciCentre to study plant immune receptor networks. Pls RT. jobs.cam.ac.uk/job/37506/ @plantsci @niabgroup
English
Samuel W. Failor retweetledi
1. New work combining 2-photon with in situ transcriptomics. With @BugeonS, @MarioDipoppa, Izzie Prankerd, Anne Ritoux, Dimitris Nicoloutsopoulos, @davidorme16, Max Shinn, Han Peng, @hamishforrest, Aiste Viduolyte, @charureddy, @yohisogai, @MatteoCarandini biorxiv.org/content/10.110…
English
Samuel W. Failor retweetledi
Front cover! Beautiful image by @mpcontreras4 @PLOSBiology @kamounlab @AdamWu9527 @HiroakiAdachi3 journals.plos.org/plosbiology/is…
Featuring our work on NRC suppressors! #Teamwork #TeamNRC
English
Samuel W. Failor retweetledi
Sad to be moving on from @KamounLab but very excited to start a new chapter at @CropSciCentre @plantsci @niabgroup with some amazing colleagues @Seb_EvdA Natasha @gilesedoldroyd Uta @harryinashed.
Crop Science Centre@CropSciCentre
👋👋🎉We are really excited to welcome three new fellows to the Crop Science Centre. Read about each of these fellows and their research by following this link cropsciencecentre.org/news/three-new… @lderevnina @Seb_EvdA
English
Samuel W. Failor retweetledi
@byronprecio @MatteoCarandini @kennethd_harris Good question! I think it is unlikely since we found more sparsening for the task-informative orientations, 45 and 90, than for the 68 distractor. With that said, we plan to look at the effect of passive task session replays on stimulus responses soon.
English
@sfailor @MatteoCarandini @kennethd_harris Do you think the mice actually need to be trained? Or, could passive exposure lead to a similar result?
English
Fascinated by learning, vision, and/or neural population codes? See our new paper, which shows that training on a 2AFC visual task sparsens and orthogonalizes the representations of task-informative stimuli in V1. w/ @MatteoCarandini and @kennethd_harris
biorxiv.org/content/10.110…
GIF
English
Samuel W. Failor retweetledi
1. How does learning affect population codes in visual cortex? We asked this in a new study with @sfailor and @MatteoCarandini. Looking at single cells, the results are perplexing. Looking at populations, they make sense. Now on bioRxiv: doi.org/10.1101/2021.0…
English