Alaister Moull

know that I was, too. I remember it all too well.

[Alaister Moull @moulla]

The answer was yes

[Alaister Moull @moulla]

Can Waren get them to bring back Verdansk?

[Beth Rigby @BethRigby]

ANALYSIS: The treatment of Jess Phllips over recent days tells me all I need to know about the epidemic of misogyny, sexual abuse and violence against women and girls that still plagues our culture.

The domestic violence campaigner-turned politician, who has spent her career fighting for victims, has found herself the subject of abuse on an industrial scale over the past week that has put her in danger, and, in dark moments, left her wondering whether she should give up frontline politics for good and go back to the women’s hostels where her journey to champion vulnerable women and girls began.

Outspoken and a women’s campaigner, Phillips has long been a lightning rod, but when the world’s richest man, who owns a social media platform with 211m followers, starts trolling you as a “rape genocide apologist”, complicit in a what he claims is a ‘cover-up’ of the most disgusting and sickening abuse, that’s a different order of attention, and danger. This week the female politician charged with trying to protect the actual victims of these unspeakable crimes becames subject to an avalanche of abuse - and threats - herself. It was undoubtedly horrific for Phillips, who tells me she felt physically sick and hunted as the tweets came raining down. And as everyone piled in with their outrage and indignation, where were the voices of the actual victims themselves?

It has been so vicious and noisy and fraught as the very serious matter of grooming gangs and the exploitation, rape and torture of young victims turned into a political battleground of finger pointing and point scoring.

You can read the full piece here > https://t.co/9WX0gGQECo

[Tim O'Reilly @timoreilly]

This is true. Algorithmic feeds are the essence of enshittification, what I've called "attention rents" that prioritize the goals of the platform over the goals of the customer.

[Alaister Moull @moulla]

IHT is a tax on familial dysfunction.

[Sam Bowman @s8mb]

I'm amazed that inheritance tax is so popular with UK elites. It strikes me as being really appalling – distortionary, anti-investment, invasive, and undermining of one of the most basic motivations humans have, which is to provide for our children.

[Alaister Moull @moulla]

@ReaperCapital Yeah, and they should absolutely present value the efficiency gains over a 20 year horizon and debt fund and all cash payout.

[Alaister Moull @moulla]

@rjrodger Helping LLM's is the human-cyborg way. We all adapt to getting Alexa to do what we want with 5 nines accuracy, by learning its strengths and weaknesses and complementing them. Those who make LLMs work for their business need to take thr same approach.

[Alaister Moull @moulla]

@jeuasommenulle Bespoke monorail expansion

[Alaister Moull @moulla]

@rjrodger 🙃

[Alaister Moull @moulla]

One hopes it will continue to feel like not living in Chernobyl

[Philipp Schmid @_philschmid]

Is that how living in the EU will feel in the Age of AI? 🤔 @AIatMeta's new multimodal Llama, @Apple Intelligence, or @OpenAI's Advanced Voice Mode of ChatGPT are all currently restricted in the EU. There's a growing concern that this could lead to "two-speed AI" - more advanced AI for the rest of the world and less capable AI for the EU.

While regulations aim to protect, they can also hinder innovation—we need to quickly find a balance to ensure progress isn't stalled. 🇪🇺

[Alaister Moull @moulla]

Skynet would want nuclear power too

[Alaister Moull @moulla]

Is this the first steps on AI's energy problem?

[Niko McCarty. @NikoMcCarty]

Mixtures of engineered bacteria were able to:

- Identify if a number is prime
- Check if a letter in a string is a vowel
- Determine the max number of pieces of a pie obtained from n straight cuts.

Answers are printed by expressing fluorescent proteins in different patterns.

[Hamel Husain @HamelHusain]

#1 Question I get as an AI consultant: am I using the best frameworks, embeddings, agent architecture, and tools?

#1 Question ppl should be asking: “Do I have metrics to track performance across various categories? Am I looking at the data? Do I know what’s failing?”

[Alaister Moull @moulla]

Getting infinite monkeys to produce the works of Shakespeare is easy. Getting them to produce something as good as the works of Shakespeare is much harder

[Alaister Moull @moulla]

Actions => Consequences

[Andrej Karpathy @karpathy]

It's a bit sad and confusing that LLMs ("Large Language Models") have little to do with language; It's just historical. They are highly general purpose technology for statistical modeling of token streams. A better name would be Autoregressive Transformers or something.

They don't care if the tokens happen to represent little text chunks. It could just as well be little image patches, audio chunks, action choices, molecules, or whatever. If you can reduce your problem to that of modeling token streams (for any arbitrary vocabulary of some set of discrete tokens), you can "throw an LLM at it".

Actually, as the LLM stack becomes more and more mature, we may see a convergence of a large number of problems into this modeling paradigm. That is, the problem is fixed at that of "next token prediction" with an LLM, it's just the usage/meaning of the tokens that changes per domain.

If that is the case, it's also possible that deep learning frameworks (e.g. PyTorch and friends) are way too general for what most problems want to look like over time. What's up with thousands of ops and layers that you can reconfigure arbitrarily if 80% of problems just want to use an LLM?

I don't think this is true but I think it's half true.

[Noam Brown @polynoamial]

Today, I’m excited to share with you all the fruit of our effort at @OpenAI to create AI models capable of truly general reasoning: OpenAI's new o1 model series! (aka 🍓) Let me explain 🧵 1/

[Alaister Moull @moulla]

@rshotton If on average there are 2 ppl in each car, how are 10 ppl dying with increased weight of the other car? Is it perhaps that the other car is a bus?

[Michael @michael_on_x_]

Crowdstrike fix. May the force be with you. Always.

[Alaister Moull @moulla]

Ryanair "Boarding now"

[Buck Shlegeris @bshlgrs]

Ryan's approach involves a long, carefully-crafted few-shot prompt that he uses to generate many possible Python programs to implement the transformations. He generates ~5k guesses, selects the best ones using the examples, then has a debugging step.