Jacky Kwok ✓

@jackyk02

Stanford CS PhD | Berkeley EECS

3 378Followers

1 000Following

130Posts total

2MViews on collected posts

โพสต์ล่าสุด

Jacky Kwok ✓@jackyk0218 Aug

@jackyk02 This is the argument for cheap models finally paying off — you're not paying for one smart answer, you're buying the right to be wrong a dozen times and still win. At 11x cheaper, verification stops being a luxury and becomes a loop you can run forever. The frontier isn 15.2K views · 124 likes · 4 reposts · 5 replies Open on X →

Jacky Kwok ✓@jackyk0218 Aug

@jackyk02 Can we actually use different models as verifiers? Like using a local Deepseek-V4-Flash to verify the GLM-5.3 API? If that works, it would be a huge leap forward. 19.8K views · 115 likes · 0 reposts · 10 replies Open on X →

Jacky Kwok ✓@jackyk0217 Aug

@jackyk02 this is actually insane 4.5K views · 21 likes · 0 reposts · 1 replies Open on X →

Jacky Kwok ✓@jackyk0217 Aug

Scaling self-verification with DeepSeek V4 Flash beats Claude Fable 5 on Terminal-Bench 2.1, while being 11x cheaper 💰 As open-source models become more capable, they can now generate large numbers of high-quality candidate solutions and verify their own outputs at very low http 1M views · 3.1K likes · 400 reposts · 154 replies Open on X →

Jacky Kwok ✓@jackyk028 Jul

How can we extract richer signals from AI Feedback? Introducing LLM-as-a-Verifier✨— a simple verification scaling framework that achieves SOTA on agentic benchmarks 🚀 The key idea: - Use fine-grained scoring granularity (e.g., 1-20 instead of the standard 1-5 scale) - Take http 997.3K views · 553 likes · 73 reposts · 24 replies Open on X →

เทียบกับบัญชีขนาดเดียวกัน

4 โพสต์จาก 90 วันที่ผ่านมา เทียบกับช่วง under 10K ผู้ติดตาม แสดงในวงกว้าง แต่มีผู้ชมตอบสนองน้อย.

ยอดดูมัธยฐาน17 509บัญชีนี้4 175ค่ามัธยฐานของ under 10K

การเข้าถึง, %518.32%บัญชีนี้250.82%ค่ามัธยฐานของ under 10K

การมีส่วนร่วม, %0.56%บัญชีนี้1.41%ค่ามัธยฐานของ under 10K

ตัวชี้วัด	บัญชีนี้	ค่ามัธยฐานของ under 10K	อัตราส่วน
ยอดดูมัธยฐานต่อโพสต์	17 509	4 175	4.19×
การเข้าถึง (ยอดดู ÷ ผู้ติดตาม)	5.2× audience	2.5× audience	2.07×
อัตราการมีส่วนร่วม	0.56%	1.41%	0.40×

บัญชีอื่นในช่วงนี้ → เปรียบเทียบกับบัญชีอื่น → ค่าอ้างอิงเหล่านี้คำนวณอย่างไร →

Growth & engagement

How the posts we collected actually performed: views and reaction rate post by post, what the audience did with them, and where the follower count goes.

Views per post

997.3K8 Jul

1M17 Aug

4.5K

19.8K18 Aug

15.2K

Last 5 collected posts, oldest on the left. The scale is logarithmic: one post can outrun the rest a hundred times over.

Engagement rate per post

0.07%8 Jul

0.37%17 Aug

0.49%

0.63%18 Aug

0.87%

Reactions — likes, reposts, replies and quotes — divided by views. Median for under 10K accounts is 1.41%.

What the audience does

Likes46.4%3 924 in total

Reposts5.6%477 in total

Replies2.3%194 in total

Quotes1.6%133 in total

Bookmarks44.1%3 733 in total

Share of every reaction we collected for this account. Replies mean argument, reposts mean endorsement, bookmarks mean the post was worth keeping.

The follower curve appears once this account has two daily snapshots — we take one a day, and this one is on its first.