Posted 5 single-line, question-led comments on posts in Rahil's space (AI in education, EdTech, founders, AI tooling, AI builders). All different creators, no overlap with prior runs. Pushed for higher-reach and stronger EdTech adjacency this time.
Author: Alicia Harper, EdM (1st connection, Superintendent / Founder / School Turnaround Strategist, SXSW EDU 2026 Speaker)
Post: "I increased proficiency from 35% to 90% by shifting one thing about our Instructional Walkthroughs" — picking ONE focus, hunting for it in every classroom, feeding trends into staff PD. 304 reactions, 257 comments.
Theme: EdTech / classroom instruction (direct Slate adjacency, Alicia is already a contact)
Comment posted:
Narrowing to one instructional focus that everyone hunts for in every classroom is what turns a walkthrough from observation into actual instruction change, did teachers initially resist the narrowness before the trend data won them over?
Author: Md Zeeshan Jawed (1st connection, Developer Relations APAC @ Eigen Labs)
Post: Built Thesis, a research assistant using Anakin's Holocron Wire (114 parallel sources) + Darkbloom by EigenCloud (Qwen3.5 122B on hardware-attested Apple Silicon) for decentralized inference. Server-Sent Events end-to-end, sub-second per source. 25 reactions.
Theme: AI builder / parallel inference / decentralized compute
Comment posted:
Parallel fan-out across 114 sources with AI source selection rather than keyword match is the part most research agents still get wrong, how did you decide when to stop adding sources versus tighten the intent classifier?
Author: Laszlo Bock (3rd+ connection, 3x founder, ex-Google SVP People Operations, NYT bestselling author of "Work Rules!", Berkeley faculty)
Post: "Becoming a doctor takes 14 years. Becoming a master plumber takes 8. Getting a great job takes an afternoon" — 30 questions × 2 answers × 3 reps = 180 reps interview prep framework. 497 reactions, 73 comments, 31 reposts.
Theme: Founder career wisdom / preparation framework
Comment posted:
The 180 reps math reframes interview prep from vague effort into a finite engineering problem, do you find founders preparing for fundraising pitches benefit from the same predict-write-rehearse loop or does investor unpredictability break it?