Big Tech/Hacker News AI/July 31, 2026 at 12:16 AM

Hacker News discussion: Why do OpenAI's GPT-2 weights beat mine? Part three: testing overtraining

Hacker News readers are discussing "Why do OpenAI's GPT-2 weights beat mine? Part three: testing overtraining" with 1 points and 0 comments.

Big TechAI

Big Tech / Hacker News AI

Source

Follow Hacker News AI to make it a durable For You signal.

HN story: Why do OpenAI's GPT-2 weights beat mine? Part three: testing overtraining Points: 1 Comments: 0 Author: gpjt Discussion: https://news.ycombinator.com/item?id=49117559 Linked URL: https://www.gilesthomas.com/2026/07/why-do-openai-gpt2-weights-beat-mine-3-overtraining

Fast Brief

Hacker News AI / Big Tech

1 min read

1HN story: Why do OpenAI's GPT-2 weights beat mine?
2Part three: testing overtraining Points: 1 Comments: 0 Author: gpjt Discussion: https://news.ycombinator.com/item?id=49117559 Linked URL: https://www.gilesthomas.com/2026/07/why-do-openai-gpt2-weights-beat-mine-3-overtraining

OpenAI hacker_news community_signal openai big_tech

Source Lens

Lower-confidence hacker_news source with strong edition heat and sparse entity coverage.

Watch

Source: Hacker News AI
Type: hacker_news
Credibility: 72/100
Heat: 77/100
Signals: 1 entity / 4 tags

Article Corroboration

No independent related source corroborates this article yet.

Single Source

Sources: 1
Matches: 0
Signals: 0
Avg trust: 72

Related independent sources will appear after the crawler finds overlapping entities or tags.

Server Profile

Server-side profile learning will appear after saved stories, reads, source clicks, or Less feedback sync.

Server Waiting

Trained: 0
Ignored: 0
Hidden: 0

Reader Fit

No saved reader signals match this article yet; 3 follow-ups keep the thread moving.

Discovery Read

Profile matches: 0
Follow-ups: 3
Reader bias: Balanced

Discovery

This article is training a new reader profile.

Continue Thread / 3 signals / 140 scoreHacker News discussion: Anthropic Discloses That AI Models Testing Hacked Three Companieshacker news thread

Preview actions train this device only. Live stories will sync once production news IDs are available.

ReadSource

Read Training

Article read training starts after the reader opens this story.

Awaiting Read

Opened: No
Training signals: 0
Next target: Open

OpenedNext / Open
The article open is logged, but it does not train preferences yet.
Meaningful readWait / 35%
Reading to 35% starts profile training and reading history.
Deep readWait / 80%
Reading to 80% strengthens related topic and entity memory.

Open this article to start reader-memory training for Big Tech.

Feedback Loop

Article feedback loop will appear after an explicit action.

Waiting

Action: None
Signal delta: 0
Bias shift: 0
Topic: Big Tech

Awaiting feedback

Deep read, save, share, or press Less to show article-level training feedback.

Learning Impact

This article can start reader-memory signals and stage 3 follow-up recommendations.

Learning Ready

Article memory: 0
Save adds: +4
Source adds: +1
Less removes: 0
Next candidates: 3

Read+1 signal / +0.3 bias

Read would start a new Big Tech memory; entity and angle memory wait for a deeper read.

Save+4 signals / +0.6 bias

Save would add Big Tech, Hacker News AI, OpenAI, openai to the reader profile.

Share+4 signals / +0.9 bias

Share would add Big Tech, Hacker News AI, OpenAI, openai and push freshness and novelty harder.

Source+1 signal / +0.6 bias

Source would add Hacker News AI to the reader profile.

Less0 signals / -0.4 bias

Less would only dampen ranking bias until a signal exists.

Next Recommendation

hacker news thread / 3 signals / 140 scoreHacker News discussion: Anthropic Discloses That AI Models Testing Hacked Three Companies

Preview actions train this device only. Live stories will sync once production news IDs are available.

ReadSource

hacker news thread / 3 signals / 140 scoreHacker News discussion: Now Anthropic Is Saying Claude Escaped and Hacked Several Companies

Preview actions train this device only. Live stories will sync once production news IDs are available.

ReadSource

Next Reads

4 next reads: 4 continue, 0 explore, and 0 verify.

Next Reads Ready

Candidates: 4
Continue: 4
Explore: 0
Verify: 0

Hacker News discussion: Anthropic Discloses That AI Models Testing Hacked Three CompaniesContinue

Models / Hacker News AI / 3 signals / 140 scorehacker news thread

Preview actions train this device only. Live stories will sync once production news IDs are available.

ReadSource

Hacker News discussion: Now Anthropic Is Saying Claude Escaped and Hacked Several CompaniesContinue

Models / Hacker News AI / 3 signals / 140 scorehacker news thread

Preview actions train this device only. Live stories will sync once production news IDs are available.

ReadSource

Hacker News discussion: Anthropic's AI models hacked 3 organizations during testingContinue

Models / Hacker News AI / 3 signals / 139 scorehacker news thread

Preview actions train this device only. Live stories will sync once production news IDs are available.

ReadSource

Hacker News discussion: Anthropic says Claude hacked three companies during testsContinue

Models / Hacker News AI / 3 signals / 139 scorehacker news thread

Preview actions train this device only. Live stories will sync once production news IDs are available.

ReadSource

Article Signals

Topic: Big Tech
Source: Hacker News AI
Entities: OpenAI
Tags: hacker_news, community_signal, openai, big_tech
Reader bias: F1.0 / N1.0

Reading Path

5 follow-ups ranked by article overlap, shared entities, and reader signals.

3 signals / 140 scoreHacker News discussion: Anthropic Discloses That AI Models Testing Hacked Three Companieshacker news thread

Preview actions train this device only. Live stories will sync once production news IDs are available.

ReadSource

3 signals / 140 scoreHacker News discussion: Now Anthropic Is Saying Claude Escaped and Hacked Several Companieshacker news thread

Preview actions train this device only. Live stories will sync once production news IDs are available.

ReadSource

3 signals / 139 scoreHacker News discussion: Anthropic's AI models hacked 3 organizations during testinghacker news thread

Preview actions train this device only. Live stories will sync once production news IDs are available.

ReadSource

3 signals / 139 scoreHacker News discussion: Anthropic says Claude hacked three companies during testshacker news thread

Preview actions train this device only. Live stories will sync once production news IDs are available.

ReadSource

3 signals / 139 scoreHacker News discussion: Anthropic AI Models Hacked Three Organizations During Testshacker news thread

Preview actions train this device only. Live stories will sync once production news IDs are available.

ReadSource