Toots for bornach@masto.ai account

Written by Bornach on 2025-01-26 at 21:11

Corridor Crew call out AI slop in the LA fire video clips circulated on social media

https://youtu.be/zhdHTax1wyQ

[#]AIslop #LAfires #Fake #generativeAI

=> More informations about this toot | View the thread

Written by Bornach on 2025-01-13 at 21:21

While researching domino tilings, #Google's #AI slop bot unhelpfully makes misleading claim when I never asked for it.

Can you prove it wrong?

[#]AIslop #mathematics #maths #math #mathsodon #LLM #puzzle

=> View attached media

=> More informations about this toot | View the thread

Written by Bornach on 2024-12-31 at 08:13

[ThioJoe] uses #NotebookLM to generate an entire podcast including voice audio, not just a script, about a text file containing just the words "shart and fart"

https://youtu.be/iV6TS7Ww4Ds

The human speech synthesis is highly convincing. How much are you willing to bet #Google #DeepMind trained their #AI model using real podcasts without first asking the creators for permission?

[#]generativeAI #artificial_intelligence

=> More informations about this toot | View the thread

Written by Bornach on 2024-12-16 at 15:56

[Synthet] lists all the elements that make a song sound Christmassy

youtu.be/VUBH9u8Gx2g

[#]Christmas #song #popmusic #music #composition

=> More informations about this toot | View the thread

Written by Bornach on 2024-12-13 at 12:42

Plotting the graph to find out how many solutions there were would have misled you into thinking there were only 3

https://youtu.be/XOBKP4Hvpbo

Why did they not all show up in his plot?

[#]mathematics #mathsodon #maths #math #puzzle

=> More informations about this toot | View the thread

Written by Bornach on 2024-12-06 at 09:07

[AI Explained] assesses the technical report, release notes and benchmarks surrounding #OpenAI #ChatGPT #o1pro mode

https://youtu.be/AeMvOPkUwtQ

Better at math and programming but still fails at commonsense reasoning tasks about as often as previous models, nowhere close to #AGI. For some tasks paying Open AI $200/month for the #o1promode access gets you worse results

[#]AIhype #GenerativeAI #artificial_intelligence

=> More informations about this toot | View the thread

Written by Bornach on 2024-12-05 at 14:28

Google unleashes AI on YouTube that you cannot turn off. It generates AI slop summaries of videos and suggests replies to comments that contain many factual errors. The AI seems convinced the content creator of Franlab is a man who is afraid of technology

https://youtu.be/DteWHExa04I

[#]AI #ArtificialIntelligence #LLM #AIslop #GenerativeAI #SkyNet

=> More informations about this toot | View the thread

Written by Bornach on 2024-10-31 at 12:03

Matt Parker (not the asteroid) explains the #regex incantation that manifests prime numbers

https://youtu.be/5vbk0TwkokM

[#]mathematics

=> More informations about this toot | View the thread

Written by Bornach on 2024-10-15 at 12:48

Scam piano course advertised via Youtube linked to Scientology

https://youtu.be/BOKv8vMODuQ

[#]music #piano

=> More informations about this toot | View the thread

Written by Bornach on 2024-10-14 at 20:46

ChatGPT4-o1 can code as well as a junior developer according to [Internet of Bugs] a Youtuber known for debunking the DevinAI hype. However just don't ask it to use a niche language, the very latest version of a library, write regression tests that catch obvious bugs, or make sense of and modify megabytes of legacy source code

https://youtu.be/j0yKLumIbaM

[#]ChatGPT #chatgpt4o1preview #chatgpt4o1 #generativeAI

[#]ArtificialIntelligence

=> More informations about this toot | View the thread

Written by Bornach on 2024-10-07 at 13:03

This puzzle and its many interpretations were discussed in this [Mind Your Decisions] video

https://youtu.be/vZ_v6pyqaw4

=> More informations about this toot | View the thread

Written by Bornach on 2024-10-07 at 13:01

I have a wooden chopping board about the same size and shape as a pizza.

If it took 10 minutes to saw a board into 2 pieces, how long would it take to saw another board into 3 pieces?

[#]puzzle #maths #mathstodon #math

=> View attached media | View attached media

=> More informations about this toot | View the thread

Written by Bornach on 2024-09-05 at 13:59

The real explanation may be far more prosaic. There are plenty of forums where people challenge each other to create acrostics of different words. For example

https://coasterforce.com/forums/threads/acrostic-poems.37323/

These forums have most likely been scraped into The Pile or Common Crawl that is used to train Large Language Models. So more than enough examples for an AI to see past its tokenization.

However there are very few forums devoted to counting r's in words, or using the second letter of words to spell other words.

=> More informations about this toot | View the thread

Written by Bornach on 2024-09-05 at 13:39

The tokenization apparently didn't get in the way of Perplexity AI when asked to:

Write a sentence where the first letter of each word spells out "STRAWBERRY"

However change "first" to "second" and it completely loses the plot

=> View attached media | View attached media

=> More informations about this toot | View the thread

Written by Bornach on 2024-09-05 at 13:22

If the tokenization prevents a LLM from seeing the individual letters that spell a word, then why are all LLMs from the big AI tech companies so good at creating acrostics?

In the same session that Gemini miscounted the r's it was able to write this:

=> View attached media

=> More informations about this toot | View the thread

Written by Bornach on 2024-09-05 at 13:14

The flawed explanations typically center around the initial tokenization stage masking the letters so that all the LLM sees is a vector of numbers

https://youtube.com/shorts/7pQrMAekdn4

=> More informations about this toot | View the thread

Written by Bornach on 2024-09-05 at 13:10

Over the past month seeing several people commenting or posting an incorrect explanation for why LLM's give the wrong answer when asked: How many r's in the word "strawberry"

[#]LargeLanguageModels #LLM #GenerativeAI

=> View attached media

=> More informations about this toot | View the thread

Written by Bornach on 2024-08-30 at 17:16

How long until an AI generated book on how to forage for wild foods leads to a reader poisoning themselves?

https://youtu.be/kwp_WEdJaEk

[Atomic Shrimp] reviews two foraging books ordered from Amazon and are strongly suspected of being generated by a Large Language Model.

[#]GenerativeAI #AIslop #LLM #GPT #foraging

=> More informations about this toot | View the thread

Written by Bornach on 2024-08-18 at 13:36

Gary Marcus's AGI-24 talk looks back at his AGI-21 talk and explains why we don't yet have AGI today and how #generativeAI is a dress rehearsal for #AGI

https://youtu.be/91SK90SahHc

[#]ArtificialIntelligence #AIhype

=> More informations about this toot | View the thread

Written by Bornach on 2024-08-17 at 07:15

James Grime appears on [Another Roof] and points out all the problems with that 2014 movie on Alan Turing breaking the Enigma

https://youtu.be/hFqphtYN3So

[#]AlanTuring #TheImitationGame #JamesGrime #MovieReview #mathstodon #Mathematics

=> More informations about this toot | View the thread

=> This profile with reblog | Go to bornach@masto.ai account