edward challis

[PageIndex @PageIndexAI]

We've seen the same shift in RAG — not only for code, but for documents too — so we built PageIndex: vectorless, agentic, reasoning-based RAG, with a document tree index.
No vector DB, no chunking needed.
https://t.co/lUVuedhWYh

[UiPath @UiPath]

“UiPath is different, in that its agent approach has a higher focus on enterprise-wide agents…” @JasonTAndersen highlights 3 new capabilities that make our #agentic AI & #automation unique. https://t.co/6WERBg6RaW via @Forbes

[UiPath @UiPath]

Meet UiPath Autopilot™ for developers. 👨‍💻 Effortlessly build #automations, apps, & expressions in UiPath Studio & Apps.

[ALASTAIR CAMPBELL @campbellclaret]

They don’t. Believe me. Why do you talk such utter shite? And why are you morphing into a sovereign individual
anarcho-capitalist narcissist with fascist leanings? Some journey you’ve been on.

[edward challis @edwardchallis]

@resy your login in and reservation process is broken. Do you test it? Why do you add so many boxes and screens? You add friction to booking a restaurant rather than making it easier. Restaurants should choose another service.

[Ferenc Huszár @fhuszar]

I came to do my PhD in the UK (and stayed to eventually pay more taxes than 99% of Brits) only because my partner could move with me.

As a Cambridge academic, I am losing out on great students, top global talent, who choose Germany because they have a partner.

[UiPath @UiPath]

It’s time to innovate the way you communicate. 💬💡 Communications Mining turns every msg into actionable data, in real time. #GenerativeAI #automation

[UiPath @UiPath]

Are you finally ready to embed AI in your enterprise? Get started with our C-suite's guide to making AI work. ✅ #GenerativeAI

[Yann LeCun @ylecun]

Survey by country:
"Products and services using AI have more benefits than drawbacks"
China: yeah, AI good!
South Korea, Turkey, Brazil: Meh.
Europe: AI bad.
France, Canada, Netherlands, US: OMG, we are doomed!

From the Stanford AI Index:
https://t.co/qzGXd5oApw

[edward challis @edwardchallis]

@soundboy personally always rated Bengio and Hinton more highly than LeCun.

[Ian Hogarth @soundboy]

1/ Notable how three pioneers of deep learning ( recognised in their shared 2018 Turing award) have substantially diverged on how they assess risk from superintelligence:

[edward challis @edwardchallis]

@scott_sage @crane_vc @cobzarenco @cristicbz_ @reinferio Thanks @crane_vc . Couldn't have done it with out you.

[HENRY MORRIS @mrhenrymorris]

ATTENTION SERFS:

I’m delighted to announce The Diary of a Secret Tory. My insider account of 12 months inside the Conservative government – the scandalous affairs, the bad decision-making, Matt Hancock’s HUEL addiction… it’s going to be monumental! https://t.co/eAiQ5tcpXA

[Standard News @standardnews]

East London residents launch bid to turn concrete ex-water depot into wild swimming oasis https://t.co/fDji9XFjSP

[East London Waterworks Park 🏊🍃💚🍃🏊 @ELWaterworks]

East London Waterworks Park will provide #Hackney & #WalthamForest with free access to 2 Olympic-size #wildswimming ponds, a new park filled with #biodiversity, a make-and-repair space, open-access arts-science projects, and a #community hub. Donate here: https://t.co/FCgF5TV3iP

[CogX Festival @CogX_Festival]

@edwardchallis , CEO & Co-Founder of @reinferio , and @JohnVeichmanis, COO at @carwowuk discuss how to leverage AI and sales intelligence to create a growth flywheel. On Wednesday 15th June on our Summit Stage. #CogXFestival22

[Andrej Karpathy @karpathy]

Merely continuing to glide down the existing scaling laws with engineering alone to lower the loss (which has so far correlated with output magic of the emergent properties) is just the lower bound / worst case.

[Durk Kingma @dpkingma]

@LucaAmb I know, terrible cliche, but I just had to say it. Bayesian methods are beautiful in theory, and I had some papers in this direction, but alas, generally they are incredibly challenging to scale. (I'm also happy that people keep trying though, it's obviously a worthy goal!)

[hardmaru @hardmaru]

“Triangle with four corners” #dalle

[Quoc Le @quocleix]

For PaLM, chain of thought prompting boosted the solving rate on GSM8K from 17.9% (by regular prompting) to 58.1%. Quite a big jump!