Sign in

Martin McMillan, PhD

@martinmcmillan.bsky.social
770 followers 3.2K following 333 posts

Pharmaceutical, Data Science, & IT Consultant

PostsRepliesMedia
Reposted by Martin McMillan, PhD
Mayor Zohran Kwame Mamdani @mayor.nyc.gov · 24/09/2026
Never been to the opera before? Now’s your chance. We’re giving away 70,000 free tickets to the @metopera at the iconic Lincoln Center. Go to on.nyc.gov/opera to enter the lottery. There’s not a bad seat in the house.
684249175520
Martin McMillan, PhD @martinmcmillan.bsky.social · 01/08/2025
"But with Showrunner, you can make something shaped a bit like satire with middling chatbot sentences read out by a flat computer voice!" From pivot-to-ai.com
public-api.wordpress.com
Edward Saatchi: pay to make your own bad AI South Park episodes!
Edward Saatchi runs Fable Studio, which started in 2018 doing virtual reality. A couple of pivots later, Fable has an exciting new project called Showrunner that lets you make your own TV show epis…
1174
Martin McMillan, PhD @martinmcmillan.bsky.social · 30/07/2025
New from pivot-to-ai.com
public-api.wordpress.com
CSIRO posts AI-fake climate article with made-up quotes
Cosmos is — or was — a popular science magazine published by the CSIRO, Australia’s national science agency. David Ho is an oceanographer at the University of Hawai’i. Cosmos posted an articl…
061
Martin McMillan, PhD @martinmcmillan.bsky.social · 29/07/2025
"...violating privacy hard enough they don’t even dare offer it in Europe." New from pivot-to-ai.com
public-api.wordpress.com
Copilot Vision AI sends your data to Microsoft
Remember that much-loved Microsoft AI feature Copilot Recall that keeps screenshots of everything happening on your Windows PC? Now there’s Microsoft Copilot Vision for Windows 11, which takes what…
0176
Martin McMillan, PhD @martinmcmillan.bsky.social · 28/07/2025
What more is there to say? From pivot-to-ai.com
public-api.wordpress.com
Microsoft Copilot AI’s official avatar looks like a blob of semen
Microsoft Copilot has an avatar now! It’s a viscous pearly blob of liquid. That wobbles. And has facial expressions. I had to euphemise furiously on the YouTube version of this post — but for the b…
041
Martin McMillan, PhD @martinmcmillan.bsky.social · 28/07/2025
"Seraphinne Vallora are so successful in the AI fashion business that they'll sell you all the cool prompts you need to undercut them." From pivot-to-ai.com
public-api.wordpress.com
Seraphinne Vallora: dubious AI model hype. Again.
We talked about dubious claims of replacing fashion models with AI image generator output  back in April, when H&M made a series of actually impossible claims about AI models that they got prin…
042
Martin McMillan, PhD @martinmcmillan.bsky.social · 28/07/2025
"People feel they would be putting their job at risk if they pushed back on AI directives." From pivot-to-ai.com
public-api.wordpress.com
16% of employees pretend to use AI at work to please their boss
Tech recruiters Howdy surveyed 1,047 full-time professionals who say they use AI. 30% of survey subjects said they were managers. [Howdy] 75% of subjects were expected to use AI at work. 22% felt p…
2126
Martin McMillan, PhD @martinmcmillan.bsky.social · 28/07/2025
"Google claims the AI Overview on every search result — a frequently-wrong summary of other people’s work — sends a ton of clicks back to the original publishers they ripped off. This is false." From pivot-to-ai.com
public-api.wordpress.com
Nobody clicks past Google’s AI Overviews
Google claims the AI Overview on every search result — a frequently-wrong summary of other people’s work — sends a ton of clicks back to the original publishers they ripped off. This is false. Pew …
084
Martin McMillan, PhD @martinmcmillan.bsky.social · 28/07/2025
"This story reads like a tale for social media with no verifiable details..." From pivot-to-ai.com
public-api.wordpress.com
Huge SaaStr-Replit vibe coding disaster! — if it ever happened
UPDATE: reply from Lemkin appended at end, he says it really did happen this way Jason Lemkin of AI startup SaaStr posted a Twitter thread on Thursday 17 July, claiming he had vibe-coded an app usi…
031
Martin McMillan, PhD @martinmcmillan.bsky.social · 28/07/2025
"[C]opyright. Open source licenses rest on copyright. It’s copyrighted, and you have a specific license to use it." From pivot-to-ai.com
public-api.wordpress.com
Open source projects reject AI code over copyright concerns
I asked a couple of months ago: if AI code was such an obviously superior engineering method, why didn’t we see AI code everywhere in open source? Since then, there have been a few performative AI …
0114
Martin McMillan, PhD @martinmcmillan.bsky.social · 28/07/2025
"It can’t follow a script, it can’t follow direction, it can’t give consistent characters, the training data keeps leaking through, the sound effects are inept." From pivot-to-ai.com
public-api.wordpress.com
Google Veo fails week 4: the final faildown
When you’ve read this post, go watch the worked examples on YouTube. [YouTube] Our intrepid friend Aron Peterson, a.k.a. Shokunin Studio, is a media production guy. He follows the AI image generato…
0153
Martin McMillan, PhD @martinmcmillan.bsky.social · 28/07/2025
Send money! From pivot-to-ai.com
public-api.wordpress.com
Help fund Pivot to AI’s work — send a few dollars to keep the stories and videos coming
It’s that time again when I ask you to send money to Pivot to AI. Help keep the stories coming daily! Here is how to send money: The Patreon is here. Sign up for $5 a month today! Less th…
032
Martin McMillan, PhD @martinmcmillan.bsky.social · 28/07/2025
"Virtually every word...was deliberately taken from an earlier essay that I posted on Christmas Day 2022..." From @garymarcus.bsky.social
substack.com
What to Expect When You’re Expecting … GPT-5
7 predictions that might sound familiar
000
Martin McMillan, PhD @martinmcmillan.bsky.social · 13/07/2025
"The core notion has always been that the two main strands of AI—neural networks and symbolic manipulation—complement each other, with different strengths and weaknesses." New from @garymarcus.bsky.social
substack.com
How o3 and Grok 4 Accidentally Vindicated Neurosymbolic AI
Neurosymbolic AI is quietly winning. Here’s what that means – and why it took so long
0218
Martin McMillan, PhD @martinmcmillan.bsky.social · 11/07/2025
"[B]ut they just need to learn Cursor properly.” Which is another way of saying “it can’t be that stupid, you must be prompting it wrong.” New from pivot-to-ai.com
public-api.wordpress.com
AI coders think they’re 20% faster — but they’re actually 19% slower
Model Evaluation and Threat Research is an AI research charity that looks into the threat of AI agents! That sounds a bit AI doomsday cult, and they take funding from the AI doomsday cult organisat…
0132
Martin McMillan, PhD @martinmcmillan.bsky.social · 11/07/2025
"GenAI powered chatbots’ lack of reasoning can directly contribute to the nefarious effects of LLM grooming: the mass-production and duplication of false narratives online with the intent of manipulating LLM outputs." New from @garymarcus.bsky.social
substack.com
Bad Actors are Grooming LLMs to Produce Falsehoods
Our research shows that even the latest "reasoning" models are vulnerable
020
Martin McMillan, PhD @martinmcmillan.bsky.social · 11/07/2025
“[C]urrent AI tools actually slowed down task completion time by 19%.” New from @garymarcus.bsky.social
substack.com
BREAKING NEWS: AI coding may not be helping as much as you think
Coding has been the strongest use case. But a new study from METR just dropped.
020
Martin McMillan, PhD @martinmcmillan.bsky.social · 10/07/2025
"The AI bubble will last at least two, maybe three more years, because the venture capitalists really need it to. When, not if, the VCs and their money pipeline go home and the chatbot prices multiply by ten or more, the market for generative AI will collapse." New from pivot-to-ai.com
public-api.wordpress.com
‘AI is here to stay’ — is it, though? What do you mean, ‘stay’?
When some huge and stupid public chatbot disaster hits the news, the AI pumpers will Kramer into the mentions to say stuff like “you have to admit, AI is here to stay.” Well, no, I don’t. Not unles…
2174
Martin McMillan, PhD @martinmcmillan.bsky.social · 09/07/2025
"I compared chatbots to gambling and cocaine before, and Cursor fans are the most abject gutter krokodil addicts." From pivot-to-ai.com
public-api.wordpress.com
Cursor tries setting less money on fire — AI vibe coders outraged
Anysphere is the startup that produces Cursor, your sort-of dependable vibe coding buddy. You tell Cursor what you’d like and it spits out a complete function! This leads to some spectacular vibe c…
072
Martin McMillan, PhD @martinmcmillan.bsky.social · 09/07/2025
"So if you get called in to fix a vibe-slopchurned disaster, charge as much as you can. Then charge more than that." From pivot-to-ai.com
public-api.wordpress.com
Calling in the AI vibe-writing cleanup crew
You’ve heard of vibe coding, with, I’m sure, not a little horror. Spin the gacha and make a website! And then you have unfortunate experiences. There are people who get quietly called in to f…
0111
Martin McMillan, PhD @martinmcmillan.bsky.social · 09/07/2025
From pivot-to-ai.com
public-api.wordpress.com
Google AI Overviews: Publishers file competition complaints in UK and EU
The Independent Publishers Alliance, a group of small UK publishers, has filed a UK competition complaint and an EU antitrust complaint over Google’s AI Overviews in search results. [Press Gazette;…
021
Martin McMillan, PhD @martinmcmillan.bsky.social · 09/07/2025
"These bugs are all public information — so the AI models have almost certainly trained on the actual text of the bug and on the fix for the bug." From pivot-to-ai.com
public-api.wordpress.com
How to pass an AI coding benchmark: train on the questions
SWE-Bench Verified by OpenAI tests how well a model can solve real bugs in real Python code from GitHub. These bugs are all public information — so the AI models have almost certainly trained on th…
184
Martin McMillan, PhD @martinmcmillan.bsky.social · 09/07/2025
From @garymarcus.bsky.social
substack.com
Are LLMs starting to become sentient?
A compassionate but skeptical letter that Douglas Hofstadter wrote to one of his readers
000
Martin McMillan, PhD @martinmcmillan.bsky.social · 09/07/2025
From @garymarcus.bsky.social
substack.com
Scott Alexander’s Misleading Victory Lap
Whatever happened to steelmanning?
000
Martin McMillan, PhD @martinmcmillan.bsky.social · 07/07/2025
"... video cards ... that famously long-term asset class that solidly maintains its value over, ooh, months." New from pivot-to-ai.com
public-api.wordpress.com
Venture capital is lending money against video cards
Last year, Wall Street finance firms started making loans, using video cards as collateral  — that famously long-term asset class that solidly maintains its value over, ooh, months. Investment bank…
0164
Martin McMillan, PhD @martinmcmillan.bsky.social · 06/07/2025
"... [defines] botshit as the human use of untruthful LLM-generated content ..." Latest from @garymarcus.bsky.social
substack.com
Botshit Gone Wild
The downsides of generative AI are growing like wildfire
010
Martin McMillan, PhD @martinmcmillan.bsky.social · 06/07/2025
"A lot of AI fans are still telling Aron it can’t be that stupid, he must be prompting it wrong. But they still haven’t posted a prompt that will work reliably for other people." Latest from pivot-to-ai.com
public-api.wordpress.com
Google Veo 3 fails, week 3 — fail harder with a vengeance
When you’ve read this post, go watch the worked examples on YouTube. [YouTube] Our good friend Aron Peterson, a.k.a. Shokunin Studio, is a media production professional. He experiments with AI imag…
042
Martin McMillan, PhD @martinmcmillan.bsky.social · 06/07/2025
"... the “97%” free users are not in fact likely to start paying out their own pockets." New from pivot-to-ai.com
public-api.wordpress.com
Only 3%* of US AI users are willing to pay a penny for it
Venture capitalists Menlo Ventures have released what purports to be a survey: “2025: The State of Consumer AI”. That is, chatbots. [Menlo Ventures] The subtitle is: “AI’s Consumer Tipping Po…
073
Martin McMillan, PhD @martinmcmillan.bsky.social · 04/07/2025
"Microsoft isn’t replacing these guys with AI. It’s firing them so it can throw even more money at AI. Microsoft plans to spend $80 billion on AI infrastructure in the next year." New from pivot-to-ai.com
public-api.wordpress.com
Microsoft lays off the staff who make the money to fund AI that doesn’t
Microsoft is firing 9,000 workers across the company — 4% of all employees. Before this, they laid off over 6,000 in May and another 300 in June. [CNBC] It’s not that Microsoft is doing badly. In Q…
16116
Martin McMillan, PhD @martinmcmillan.bsky.social · 02/07/2025
"Retraction Watch heard from a reader who was cited in the book — but his cited papers didn’t ... exist. A lot of citations in the book don’t exist." New from pivot-to-ai.com
public-api.wordpress.com
$219 Springer Nature AI textbook was written with a chatbot
“Mastering Machine Learning: From Basics to Advanced” by Govindakumar Madhavan was published in April by Springer Nature. It’s $169 as an ebook, or $219 as a hardback. [Springer, archive] Retractio…
000
Martin McMillan, PhD @martinmcmillan.bsky.social · 01/07/2025
"Microsoft’s AI tools don’t work. Microsoft AI doesn’t make you more effective. Microsoft AI won’t do the job better." New from pivot-to-ai.com
public-api.wordpress.com
‘AI is no longer optional’ — Microsoft admits AI doesn’t help at work
An internal Microsoft memo has leaked. It was written by Julia Liuson, president of the Developer Division at Microsoft and GitHub. The memo tells managers to evaluate employees based on how much t…
13410
Martin McMillan, PhD @martinmcmillan.bsky.social · 01/07/2025
"Veo just fails. It will not take direction. You can’t tell it to do stuff and get even good results, let alone professionally usable results." New from pivot-to-ai.com
public-api.wordpress.com
Google Veo 3 fails, week 2: fail harder — Shokunin Studio spins the gacha
This is our second week of running Google’s Veo 3 video generator through its paces. So when you’ve read this, go and watch the worked examples. And check out last week’s. [YouTube] Our good friend…
071
Martin McMillan, PhD @martinmcmillan.bsky.social · 01/07/2025
Keep the news flowing! From pivot-to-ai.com
public-api.wordpress.com
You fund our work here. Send $5 to Pivot to AI!
It’s the end of the month, and that’s when I ask you to donate to Pivot to AI! At present, Pivot is my income. (Got sysadmin/devops work? Call me!) So your $5 a month — or any other amo…
011
Martin McMillan, PhD @martinmcmillan.bsky.social · 29/06/2025
"Everything that they do is through mimicry, rather than abstracted cognition across proper world models." Latest from @garymarcus.bsky.social
garymarcus.substack.com
Generative AI’s crippling and widespread failure to induce robust models of the world
LLM failures to reason, as documented in Apple’s Illusion of Thinking paper, are really only part of a much deeper problem
2348
Martin McMillan, PhD @martinmcmillan.bsky.social · 27/06/2025
"The companies are going nowhere. They can’t exit, they’re running out of money, they can’t get more funding." Latest from pivot-to-ai.com
public-api.wordpress.com
Pitchbook: AI venture capital resorts to mergers
Today, let’s look at two Pitchbook reports — the US VC-Backed M&A Outlook for second quarter 2025 and the 2025 US Venture Capital Outlook: Midyear Update. [Pitchbook; Pitchbook, PDF, archive; P…
0136
Martin McMillan, PhD @martinmcmillan.bsky.social · 27/06/2025
"The authors lost against Meta because they just didn’t present a strong enough case." New from pivot-to-ai.com
public-api.wordpress.com
Meta beats Kadrey, AI training was fair use — what this means
Meta has won against Kadrey et al — the authors’ case against Meta training their Llama LLM on the authors’ books, including on pirate copies. [Order, PDF; case docket] Both sides brought motions f…
031
Martin McMillan, PhD @martinmcmillan.bsky.social · 27/06/2025
"There is not a human on earth who wants to talk to a lying chatbot about something that’s costing them money." New from pivot-to-ai.com
public-api.wordpress.com
Verizon’s new AI chatbot — customer disservice
Much beloved US phone company Verizon has updated its My Verizon app to include 24/7 customer service from AI! It’s got a chatbot based on Google Gemini. [Verizon] Verizon tells The Verge: [Verge] …
0172
Martin McMillan, PhD @martinmcmillan.bsky.social · 27/06/2025
New from @garymarcus.bsky.social
substack.com
Image generation: Still crazy after all these years
Three years ago Google promised us “deep language understanding” for image generation. I don’t think we are there yet.
010
Martin McMillan, PhD @martinmcmillan.bsky.social · 24/06/2025
"There is no carveout, however, from the Copyright Act for AI companies." Latest from pivot-to-ai.com
public-api.wordpress.com
Anthropic AI wins broad fair use for training! — but not on pirated books
One of the many copyright cases in progress against chatbot vendor Anthropic is Bartz v. Anthropic, where several authors brought a class action suit in August 2024 for violating their copyrights w…
011
Reposted by Martin McMillan, PhD
Matt Siegel @mattsiegel.bsky.social · 22/06/2025
@irisvanrooij.bsky.social has shown you can't machine-learn your way to human level intelligence; with actual mathematical proofs
033
Martin McMillan, PhD @martinmcmillan.bsky.social · 22/06/2025
"But none of these companies have any reasonable prospects ..." New from pivot-to-ai.com
public-api.wordpress.com
Nvidia gets into money-burning AI startups
When our glorious AI bubble deflates, and we contemplate the wreckage of so many companies that did nothing but turn venture capital billions into carbon dioxide, there’s one company that wil…
022
Martin McMillan, PhD @martinmcmillan.bsky.social · 22/06/2025
"[D]eploying LLMs as agents is an accident waiting to happen." Latest from @garymarcus.bsky.social
substack.com
LLMs: Dishonest, unpredictable and potentially dangerous.
Will we ever be able to control them?
22410
Martin McMillan, PhD @martinmcmillan.bsky.social · 22/06/2025
"... Veo’s output is defective in all sorts of ways." Latest from pivot-to-ai.com
public-api.wordpress.com
We test Google Veo: impressive demo, unusable results
Over on YouTube, we run Google’s Veo 3 video generator through its paces. So when you’ve read this, go and watch the worked examples. [YouTube] Aron Peterson is a media production professional. He’…
032
Martin McMillan, PhD @martinmcmillan.bsky.social · 21/06/2025
"He couldn't get Grok to align with his own personal beliefs so he is going to rewrite history to make it conform to his views." Latest from @garymarcus.bsky.social
substack.com
1984, but with LLM’s
“Who controls the past,’ ran the Party slogan, ‘controls the future: who controls the present controls the past.’” - 1984, George Orwell
313330
Martin McMillan, PhD @martinmcmillan.bsky.social · 21/06/2025
"Whenever there’s a sale, the majority of eligible employees sell up. Nothing like confidence, eh?" Latest from pivot-to-ai.com
public-api.wordpress.com
AI venture capital resorts to feeding the secondary market
Headlines about AI companies paying “billions” for other AI companies mostly don’t mean real money, but equity in private companies with imaginary valuations. But sometimes, people want actual mone…
041
Martin McMillan, PhD @martinmcmillan.bsky.social · 21/06/2025
"[T]he volunteers would work for free to improve Google’s bot." New from pivot-to-ai.com
public-api.wordpress.com
Google bribes iNaturalist to use generative AI — volunteers quit in outrage
iNaturalist is a website that crowdsources pictures of plants and animals to help identify species. Its tagline is “A Community for Naturalists.” iNaturalist is administered by its own small charit…
0113
Martin McMillan, PhD @martinmcmillan.bsky.social · 18/06/2025
New from pivot-to-ai.com
public-api.wordpress.com
OpenAI needs its $40 billion funding deal — so it’s threatening Microsoft!
OpenAI must finish its transition from a weird charity to some sort of for-profit company. Then it can get $40 billion from SoftBank! OpenAI absolutely needs to get the stakeholders on side — parti…
051
Martin McMillan, PhD @martinmcmillan.bsky.social · 18/06/2025
New from @garymarcus.bsky.social
substack.com
Five quick updates about that Apple reasoning paper that people can’t stop talking about
The most popular rebuttal was … a joke. Literally.
000
Martin McMillan, PhD @martinmcmillan.bsky.social · 17/06/2025
"ChatGPT makes you stupid." New from pivot-to-ai.com
public-api.wordpress.com
Bad brainwaves: ChatGPT makes you stupid
In a new study, MIT Media Lab measured 55 people over four months on how well they could write an essay — either with ChatGPT, with a search engine, or just their unassisted brain. The researchers …
042
Martin McMillan, PhD @martinmcmillan.bsky.social · 15/06/2025
"Nothing in this system is separated from anything else." New from pivot-to-ai.com
public-api.wordpress.com
EchoLeak — send an email, extract secret info from Microsoft Office 365 Copilot AI
You can’t put data into an LLM and then filter the output to block some of it. Once it’s in the model, it’s in the model. If you put confidential business information into Microsoft 365 Copilot, an…
061