18

Sycophancy is the first LLM "dark pattern"

3 months ago

from seangoedecke.com RSS feed in AI

People have been making fun of OpenAI models for being overly sycophantic for months now. I even wrote a post advising users to pretend that their...

19

Deplatforming: AI edition

3 months ago

from Strange Loop Canon in AI

“This would feel like getting stabbed in the heart.”

16

Worries About AI Are Usually Complements Not Substitutes

3 months ago

from Don't Worry About the Vase in AI

A common claim is that concern about [X] ‘distracts’ from concern about [Y].

18

World Emulation via Neural Network

3 months ago

from Made by Ollin in AI

20

Video Friday: High Mobility Robots for Logistics

3 months ago

from IEEE Spectrum in AI

Video Friday is your weekly selection of awesome robotics videos, collected by your friends at IEEE Spectrum robotics. We also post a weekly calendar...

18

AI Roundup 115: US v. Google

3 months ago

from Artificial Ignorance in AI

April 25, 2025.

17

New adventures in AI hype: “Our language models are so ‘conscious’ we need to give them rights”

3 months ago

from Marcus on AI in AI

The return of nonsense on stilts, three years later

16

The OpenAI house style is exhausting

3 months ago

from seangoedecke.com RSS feed in AI

I was reading this Reddit post when I noticed a pattern: a few times now I’ve seen a negative Reddit comment that to me just screamed “written by...

43

AI #113: The o3 Era Begins

3 months ago

from Don't Worry About the Vase in AI

Enjoy it while it lasts.

40

OpenAI’s dirty December o3 demo doesn’t readily replicate

3 months ago

from Marcus on AI in AI

Don’t believe everything you see

44

o3 Is a Lying Liar

3 months ago

from Don't Worry About the Vase in AI

I love o3.

37

New Results of State-of-the-art LLMs on 4 Political Orientation Tests

3 months ago

from Rozado’s Visual Analytics in AI

One model appears closer to the center than the rest

35

You Better Mechanize

3 months ago

from Don't Worry About the Vase in AI

Or you had better not.

37

What LLMs Will Do To Jobs: All You Need is an Oracle

3 months ago

from Weighty Thoughts in AI

LLMs are Mainly Tools That Enhance Experts

31

Crime and Punishment #1

3 months ago

from Don't Worry About the Vase in AI

This seemed like a good next topic to spin off from monthlies and make into its own occasional series.

34

When you should lie to the language model

3 months ago

from seangoedecke.com RSS feed in AI

Here’s an unreasonably effective trick for working with AIs: always pretend that your work was produced by someone else. The problem is that…

49

On Jagged AGI: o3, Gemini 2.5, and everything after

3 months ago

from One Useful Thing in AI

New models and new thresholds

39

Is using AI wrong? A review of six popular anti-AI arguments

3 months ago

from seangoedecke.com RSS feed in AI

Some people really, really don’t like AI. Broadly speaking, being anti-AI is a popular left-wing position: AI is cringe, it’s plagiarism, it…

34

o3 Will Use Its Tools For You

3 months ago

from Don't Worry About the Vase in AI

OpenAI has finally introduced us to the full o3 along with o4-mini.

46

Video Friday: Robot Boxing

3 months ago

from IEEE Spectrum in AI

Video Friday is your weekly selection of awesome robotics videos, collected by your friends at IEEE Spectrum robotics. We also post a weekly calendar...

35

ML for SWEs 7: Eval-driven Model Development?

3 months ago

from Society's Backend: Machine Learning for Software Engineers in AI

Machine learning for software engineers 4-18-25

14

ML for SWEs 7: Eval-driven Model Development?

3 months ago

from Machine Learning for Software Engineers in AI

Machine learning for software engineers 4-18-25

29

AI Roundup 114: One of those weeks

3 months ago

from Artificial Ignorance in AI

April 18, 2024.

42

OpenAI’s o3 and Tyler Cowen’s Misguided AGI Fantasy

3 months ago

from Marcus on AI in AI

AI can only improve if its limits as well as its strengths are faced honestly

37

The MCP Revolution

3 months ago

from Artificial Ignorance in AI

Why this open standard is becoming essential infrastructure for AI agents.

27

AI #112: Release the Everything

3 months ago

from Don't Worry About the Vase in AI

OpenAI has upgraded its entire suite of models.

44

The Future of AI and Robotics Is Being Led by Amazon’s Next-Gen Warehouses

3 months ago

from IEEE Spectrum in AI

This is a sponsored article brought to you by Amazon. The cutting edge of robotics and artificial intelligence (AI) doesn’t occur just at NASA, or one...

32

GPT-4.1 Is a Mini Upgrade

3 months ago

from Don't Worry About the Vase in AI

Yesterday’s news alert, nevertheless: The verdict is in.

41

”Those claiming we’re mere months away from AI agents replacing most programmers” should think again

3 months ago

from Marcus on AI in AI

AI agents will change the world. But not this year.

29

How much does a 10 million token context window actually cost?

3 months ago

from Society's Backend: Machine Learning for Software Engineers in AI

Some back-of-the-napkin math for Meta's Llama 4 Scout

13

How much does a 10 million token context window actually cost?

3 months ago

from Machine Learning for Software Engineers in AI

Some back-of-the-napkin math for Meta's Llama 4 Scout

37

Altman vs Musk cage match

3 months ago

from Marcus on AI in AI

Uh oh. Now Altman wants to build a social media company, to compete with X. Who should we root for?

32

OpenAI #13: Altman at TED and OpenAI Cutting Corners on Safety Testing

3 months ago

from Don't Worry About the Vase in AI

Three big OpenAI news items this week were the FT article describing the cutting of corners on safety testing, the OpenAI former employee amicus...

48

AI as Normal Technology

3 months ago

from AI Snake Oil in AI

A new paper that we will expand into our next book

31

Sam Altman’s attitude problem

3 months ago

from Marcus on AI in AI

What happened to the man who once told the US Senate “creators deserve control over how their creations are used”?

42

A practical guide to coding securely with LLMs

3 months ago

from seangoedecke.com RSS feed in AI

Writing code with LLMs is fundamentally different from other ways of programming. LLMs are often non-deterministic and always unpredictable…

28

Monthly Roundup #29: April 2025

3 months ago

from Don't Worry About the Vase in AI

In Monthly Roundup #28 I made clear I intend to leave the Trump administration out of my monthly roundups, for both better and worse, outside of my...

36

Why is lmarena.ai dominated by slop?

3 months ago

from seangoedecke.com RSS feed in AI

When LMSYS (aka LMArena, aka Chatbot Arena) first blew up, I thought it was the best way possible of determining which LLM really was the…

39

guitars and javascript - the case for jank

3 months ago

from Solving the decision problem in AI

worse is better (and that’s why i still write javascript)

34

Software engineering under the spotlight

3 months ago

from seangoedecke.com RSS feed in AI

Think of a tech company as a giant, dimly-lit factory. Work goes on throughout the factory as components shuffle back and forth, and…

34

ML for SWEs 6: RAG Is Not Dead

3 months ago

from Society's Backend: Machine Learning for Software Engineers in AI

Machine learning for software engineers 4-11-25

16

ML for SWEs 6: RAG Is Not Dead

3 months ago

from Machine Learning for Software Engineers in AI

Machine learning for software engineers 4-11-25

35

One giant leap towards authoritarian rule in the United States

3 months ago

from Marcus on AI in AI

In the midst of the tariffs, an even bigger story is brewing: Trump is attempting to take over American Universities.

36

AI Roundup 113: Liberation Day

3 months ago

from Artificial Ignorance in AI

April 11, 2025.

45

Video Friday: Tiny Robot Bug Hops and Jumps

3 months ago

from IEEE Spectrum in AI

Video Friday is your weekly selection of awesome robotics videos, collected by your friends at IEEE Spectrum robotics. We also post a weekly calendar...

40

On Google's Safety Plan

3 months ago

from Don't Worry About the Vase in AI

Google Lays Out Its Safety Plans

40

Defending against Prompt Injection with Structured Queries (StruQ) and Preference Optimization (SecAlign)

3 months ago

from The Berkeley Artificial Intelligence Research Blog in AI

Recent advances in Large Language Models (LLMs) enable exciting LLM-integrated applications. However, as LLMs have improved, so have the attacks...

40

Poor ROI for GenAI

4 months ago

from Marcus on AI in AI

61% in a recent survey report no ROI or disappointing ROI.

40

Demonstrating Kelly Betting with Chips

4 months ago

from Win Vector LLC in AI

I have a new video demonstrating the Kelly Can’t Fail betting strategy. The idea is: this is a classroom appropriate tool for discussing allocating...

34

AI #111: Giving Us Pause

4 months ago

from Don't Worry About the Vase in AI

Events in AI don’t stop merely because of a trade war, partially paused or otherwise.