Believer @bayesbeliever

PhD candidate in Machine Learning 🎓 Studying reinforcement learning + causal inference Exploring how AI can accelerate real scientific discovery Joined March 2025

Tweets

227
Followers

7
Following

71
Likes

877

Novix @Novix_science

3 days ago

PhD workflow in 4 steps: 1. Find 200 PDFs 2. Format 100 spreadsheets 3. Write 50 drafts 4. Make 1 discovery (if lucky) It’s not the science that’s hard. It’s the workflow

0 12 13 221 1

Most PhD students confuse reading with research But progress only happens when reading → synthesis → hypothesis A better workflow: 1. Don’t just highlight; extract the claims 2. Compare across 3+ papers, not one 3. Ask: What’s missing between them? That’s your hypothesis…

0 11 14 269 1

Novix @Novix_science

5 days ago

40% filtering papers 30% formatting data 20% analysis 10% discovery Scientists don’t need more hours They need workflows that turn time into discovery, not busywork Try novix.science

Novix @Novix_science

2 weeks ago

40% filtering papers 30% formatting data 20% analysis 10% discovery Scientists don’t need more hours They need workflows that turn time into discovery, not busywork Try novix.science

0 11 18 495 0

Download Image

0 10 15 303 0

Novix @Novix_science

4 days ago

AI agents aren’t just hype; the data shows they’re becoming the backbone of applied AI. From our latest analysis: 1. Industry share (2024): Customer service leads with 35%. Healthcare (18%) and finance (16%) are the fastest-growing application areas 2. Funding patterns: Median…

0 12 15 218 1

Download Image

Novix @Novix_science

2 weeks ago

Most AI tools stop at summaries. That’s why progress feels stuck We’re not building another summarizer We’re building an AI co-scientist From hypothesis → analysis → results → validation @Novix_science brings it all together in one place 👇

0 13 19 887 1

Download Image

Novix @Novix_science

7 days ago

Reproducibility is the foundation of science If others can’t repeat your result, it isn’t reliable knowledge Yet across fields from psychology to biology, failed replications are the norm, not the exception

Novix @Novix_science

a week ago

0 10 12 478 0

1 10 16 346 0

Novix @Novix_science

a week ago

PhD students don’t need more tools. They need an AI co-scientist

0 14 15 228 0

Novix @Novix_science

a week ago

70% of scientists report failing to reproduce results from other studies This isn’t a failure of intelligence It’s a failure of workflows that treat reproducibility as an afterthought

0 10 12 478 0

Novix @Novix_science

a week ago

If you could save 5 hours a week as a PhD student, where would you spend it?

1 13 17 168 1

Novix @Novix_science

2 weeks ago

Research cheat codes PhD students wish they knew earlier

1 13 16 212 1

Novix @Novix_science

2 weeks ago

PhD burnout doesn’t come from failed ideas It comes from drowning in repetitive tasks that make discovery feel out of reach Fix the workflow → fix the burnout Try: Novix.science

0 10 14 160 1

Novix @Novix_science

2 weeks ago

Why do PhD students spend months stuck in research? Not because they lack ideas The workflow itself wastes most of their time: • Endless PDFs → disconnected knowledge • Raw data → endless formatting • Drafting → repetitive manual work Proposed solutions: • Automate…

0 11 18 495 0

Download Image

Regrets @Regrets10x

2 weeks ago

It could be....

76 43 274 9K 16

Download Image

Novix @Novix_science

2 weeks ago

Science isn’t slow because of a lack of talent It’s slow because 80% of research time is lost to busywork, formatting, searching, cleaning data Every hour wasted there is an hour stolen from discovery Fixing that isn’t a convenience. It’s the future of science

0 12 15 199 0

Novix @Novix_science

2 weeks ago

PhD students don’t just need more hours They need better workflows Here are 10 insights on research progress, from tools to habits, that make discovery faster 👇 1. Read at scale — scan thousands of papers and extract key insights 2. Connect ideas — map patterns across…

0 11 16 806 0

Download Image

Novix @Novix_science

3 weeks ago

Ever wished you had a co-scientist? Not just to summarize papers, but to actually connect insights, analyze data, and surface new hypotheses That’s what we’re building with @Novix_science → an AI co-scientist for research Try now: novix.science