High Output AI

High Output AI

By High Output AITechnology
Download on the App Store

High Output AI episodes

  • $100M+ seed rounds for Stability & Jasper AI | EP022

    This week Tom and Elliot talk about the $225m that flew into the seed rounds for Stability AI and Jasper AI. They also touch on the rumours about OpenAI's discussions with Microsoft about further financing.


    ⏲️Timestamps⏲️

    00:00 100% more audio introduction

    01:30 The three big invesments pieces of news this week

    05:00 Two of the biggest seed rounds

    06:10 Stability AI's seed round & the universe

    12:45 Making an episode fully AI

    13:35 Detailed information on the Stability AI round

    15:00 Don't pick the hype for your PhD

    17:40 The Hedge Fund Play by Stability AI

    18:30 4000 GPUs from Stability AI

    20:40 Stability delays the release of v1.5 of Stable Diffusion and the consequences

    27:15 Elliot's spicy take🌶️ and Stavility AI's competive new future

    31:15 OpenAI's discussion with Microsoft

    35:00 The release history by OpenAI and the trend of closing

    38:45 Imagik Paper and reinventing images

    41:00 The challenges of prompt engineering

    44:00 Final words on Stability AI's round

    48:12 Jasper AI's round

    52:00 Marketing dribble just exploded 100x

    53:20 Advertising doesn't work that way

    54:30 All consumer brand's will be gone in 30 years

    59:00 The Sideways effect on the wine market

    1:01:30 Pineapples & social signalling

    1:03:20 Humans filtering out AI content

    1:04:40 Jasper's product demo

    1:06:10 Propogating errors through the internet

    1:11:00 Wrapping up


    ⚡Show Notes⚡

    - Stability AI's $101m Seed Round https://techcrunch.com/2022/10/17/stability-ai-the-startup-behind-stable-diffusion-raises-101m/

    - Jasper AI's $125m Seed Round https://techcrunch.com/2022/10/18/ai-content-platform-jasper-raises-125m-at-a-1-7b-valuation/

    - OpenAI in discussion with Microsoft for further funding https://www.theinformation.com/articles/openai-valued-at-nearly-20-billion-in-advanced-talks-with-microsoft-for-more-funding

    - Not Andrew Street... Andrew Scott https://www.imdb.com/name/nm0778831/

    - Stable Diffusion Launch https://stability.ai/blog/stable-diffusion-announcement

    - Dreamstudio https://beta.dreamstudio.ai/home

    - Ali Abdaal AI tweet https://mobile.twitter.com/AliAbdaal/status/1582835494475698176

    - Moonbeam editing tool https://www.gomoonbeam.com/

    - Emad Mostaque CEO of Stability AI https://en.wikipedia.org/wiki/Emad_Mostaque

    - Stability AI demo day https://www.youtube.com/watch?v=1Uy_8YPWrXo&t=2415s&ab_channel=RobertScoble

    - Why Stability delayed v1.5 release of Stabel Diffusion https://danieljeffries.substack.com/p/why-the-future-of-open-source-ai

    - Imagik https://imagick.ai/

    - Jasper AI Demo https://www.jasper.ai/demo

    - Ads don't work that way https://meltingasphalt.com/ads-dont-work-that-way/

    - The Sideways Effect on Merlot https://www.npr.org/sections/thesalt/2017/07/05/535038513/the-sideways-effect-how-a-wine-obsessed-film-reshaped-the-industry

    - The Myth of Spinach https://www.compoundchem.com/2018/07/17/spinach/

    -Malcolm Gladwell on memory https://www.pushkin.fm/podcasts/revisionist-history/free-brian-williams


    👋CONTACT

    Any feedback? Leave a comment or hit us on twitter

    @tom_k_white, @elliot_c_smith


    👋FRIENDSSSSSS

    Join our Discord https://discord.gg/kjKT5vp4


    #podcast #tech #ai #startup

    1 hr 13 min
  • Tesla AI Day & Data Laundering -- IN PERSON! | EP021

    For the first time in a long time Elliot and Tom are in the same room. This week we talk mostly about Tesla's AI day, as well as touching on the "Data Laundering " that is happening in the image generation world. Warning, we get pretty boozed towards the end.


    ⏲️Timestamps⏲️

    00:00 Drinking in person

    02:00 Part 1 - Tesla AI Day intro

    07:20 NERF & other Vision AI models

    14:50 Lanes languages and using attention in vision

    21:15 Data & Training capabilities behind FSD

    25:25 The infrastructure behind it all

    31:00 Tesla makes their own silicon + NVIDIA rambles

    38:35 Final comments on Tesla AI day

    43:16 Data Laundering

    46:55 OpenAI question to US Patents & Tradmark office

    55:42 Stadia died because no one trusts google


    ⚡Show Notes⚡

    Part 1 - Tesla AI Day

    - Tesla AI Day stream - https://www.youtube.com/watch?v=ODSJsviD_SU&ab_channel=Tesla

    - Radiance Fields - https://datagen.tech/guides/synthetic-data/neural-radiance-field-nerf/

    - LIDAR is doomed, Elon Musk - https://youtu.be/HM23sjhtk4Q

    - Elon Musk's perfume - https://gizmodo.com/elon-musks-100-burnt-hair-perfume-has-sold-out-1849676199

    - Russian army using chips from fridges - https://www.reddit.com/r/UkraineWarVideoReport/comments/xiqhu4/ursula_von_der_leyen_the_russian_army_is_removing/

    - Meta announces Legs - https://techcrunch.com/2022/10/11/meta-announces-legs/

    Part 2

    - Data Laundering Article - https://waxy.org/2022/09/ai-data-laundering-how-academic-and-nonprofit-researchers-shield-tech-companies-from-accountability/

    - OpenAI request for comment - https://www.uspto.gov/sites/default/files/documents/OpenAI_RFC-84-FR-58141.pdf

    - Temtem - https://store.steampowered.com/app/745920/Temtem/

    Part 3 - Google Stadia

    - Stadia died because no one trusts google - https://techcrunch.com/2022/10/01/stadia-died-because-no-one-trusts-google/

    - Player with 6k hours of Red Dead Redemption 2 loses everything https://www.cbr.com/red-dead-redemption-2-player-6000-google-stadia-hours-rockstar/


    👋CONTACT

    Any feedback? Leave a comment or hit us on twitter

    @tom_k_white, @elliot_c_smith

    #podcast #tech #ai #startup

    1 hr 3 min
  • Text to Video, New GPUs from Intel and NVIDIA's rocky future | EP020

    This week Tom and Elliot cover new models, new hardware and all the interesting things they came across this week.

    ⏱️Time stamps⏱️

    00:00 Introduction

    02:53 DALLE-2 removes their wait list

    07:42 Meta releases a paper on Make-a-video, a text to video model

    12:01 Dreamfusion's text to 3D model generation

    16:48 VTOONIFY released for converting portrait videos into caricatures

    20:32 StockAI launched and just probably has legal action coming their way

    23:33 NVIDIA Launches Cloud SaaS business

    26:53 AMD Ryzen700 CPU Launch

    29:00 New Intel GPUs

    30:48 Reducing 100k equations to 4 using AI

    35:25 NVIDIA in the valley

    41:38 EchoNet outperforms sonographers

    44:30 Learning a language by talking to AI

    47:00 Designing the perfect keyboard using AI

    49:05 Is Stadia really dead?

    55:00 Closing bants


    ⚡Show Notes⚡

    💿Model News

    - DALL-E removes their wait list… https://www.theverge.com/2022/9/28/23376328/ai-art-image-generator-dall-e-access-waitlist-scrapped

    - Meta releases a paper on a model called Make-a-video - https://ai.facebook.com/blog/generative-ai-text-to-video/

    - Dreamfusion text to 3D model https://dreamfusion3d.github.io/

    - VTOONIFY https://www.mmlab-ntu.com/project/vtoonify/

    - StockAI launched https://www.stockai.com/

    - Original tweet showing the copyright issues StockAI faces https://twitter.com/frantzfries/status/1574557008950329354

    🖥️ Hardware News

    - NVIDIA releases cloud service for metaverse creators https://nvidianews.nvidia.com/news/nvidia-launches-omniverse-cloud-services-for-building-and-operating-industrial-metaverse-applications

    - AMD new CPU launch https://arstechnica.com/gadgets/2022/09/everything-you-need-to-know-about-zen-4-socket-am5-and-amds-newest-chipsets/

    - New Intel GPUs https://www.intel.com/content/www/us/en/products/details/discrete-gpus/arc.html

    🔬 Science News

    - Researchers reduce equations from 100k to 4 https://www.simonsfoundation.org/2022/09/26/artificial-intelligence-reduces-a-100000-equation-quantum-physics-problem-to-only-four-equations/

    🪣Misc. Bucket

    - NVIDIA in the valley, Stratechery https://stratechery.com/2022/nvidia-in-the-valley/

    - EchoNet outperforming sonographers for LV function test https://www.tctmd.com/news/echonet-ai-algorithm-beats-sonographers-estimating-lv-function

    - Learning a language by talking to AI https://talk.quazel.com/chat/try

    - Stadia is officially dead https://www.theverge.com/2022/9/29/23378713/google-stadia-shutting-down-game-streaming-january-2023

    57 min
  • New GPUs, New Courses, New Models & Emerging AI | EP019

    In a new format, Tom and Elliot talk about a collection of interesting new releases from the last couple of weeks. Hope you're enjoying the show.


    ⌛Timestamps

    00:00:00 - Start

    00:03:04 - New NVIDIA GPUs

    00:15:00 - Deepmind investigates Emergent Abilities of AI

    00:30:23 - Getty Images Bans AI Content

    00:35:57 - Chal-E & Blender Textures Using AI

    00:44:20 - OpenAI Whisper

    00:46:33 - McKinsey Report on AI in China

    00:53:02 - New Fast AI Course & ACT-1 Preview


    ⚡Show Notes⚡

    New NVIDIA chip releases

    — NVIDIA 4000 series - https://www.nvidia.com/en-gb/geforce/graphics-cards/40-series/

    — Coreteks breakdown of NVIDIA performance - https://www.youtube.com/watch?v=ATS-fyQZuwA&ab_channel=Coreteks

    — Tom breaks down the new GPU perfromance - https://www.youtube.com/shorts/7raAVrdilSw

    — Ada Lovalace Architecture - https://www.nvidia.com/en-gb/geforce/ada-lovelace-architecture/

    — Ada Lovelace - https://en.wikipedia.org/wiki/Ada_Lovelace

    — Google Colab - https://colab.research.google.com/

    — Buy Elliot's coffee machine...

    — Wikitext-2 - https://paperswithcode.com/dataset/wikitext-2

    Emerging AI

    — Emerging AI by DeepMind- https://bdtechtalks.com/2022/08/22/llm-emergent-abilities/

    — Getty Images bans AI generated images - https://www.theverge.com/2022/9/21/23364696/getty-images-ai-ban-generated-artwork-illustration-copyright

    — Ian Goodfellow & GANs - https://arxiv.org/abs/1406.2661

    AI Plugins for Designer tools

    — CHARL-E - https://www.charl-e.com/

    — Blender Stable Diffusion plugin - https://carlosedubarreto.gumroad.com/l/ceb_sd

    — Stable Diffusion for Photoshop - https://www.flyingdog.de/sd/

    — Competition - https://www.reddit.com/r/technology/comments/x4swd7/a_man_won_an_art_competition_in_colorado_using/

    OpenAI Whisper

    — OpenAI Whisper - https://openai.com/blog/whisper/

    — Elliot's live dev stream of OpenAI Whisper - https://youtu.be/VlC1wMLI9NY

    McKinsey Report

    — McKinsey report on AI in China - https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-next-frontier-for-ai-in-china-could-add-600-billion-to-its-economy

    Misc.

    — Fast AI Course - https://course.fast.ai/

    — Adept.ai

    — Google assistant that could call and make bookings - https://www.youtube.com/watch?v=d40jgFZ5hXk&ab_channel=BusinessStandard

    1 hr 1 min
  • Stable Diffusion Release & Learning Physics from Video | EP018

    Tom and Elliot talk about the release of stable diffusion, some of its strengths and weaknesses and an AI model that can learn fundamental equations of real world systems from video.

    00:00 - Start

    04:14 - Topic 1 - Stable Diffusion
    27:58 - Topic 2 - Discovering Physics from Video

    ⚡Links

    — Stable Diffusion - https://stability.ai/blog/stable-diffusion-public-release
    — Automated discovery of fundamental variables hidden in experimental data - https://www.nature.com/articles/s43588-022-00281-6

    54 min
  • Amazon Buys One Medical | Podcast EP017

    Tom and Elliot talk about the recent acquisition of One Medical. Why Amazon would want it, where AI might fit into the picture and some tin hat theories of what their master plan entails.

    ⌛Timestamps

    00:00 - Start

    05:06 - AI For Marketing Misses the Mark
    40:42 - AI For Grape Yield

    ⚡Links

    — Amazon official announcement https://press.aboutamazon.com/news-releases/news-release-details/amazon-and-one-medical-sign-agreement-amazon-acquire-one-medical
    — NPR coverage of the purchase https://www.npr.org/2022/08/07/1116218333/amazon-may-own-your-doctors-office-next?t=1659968145417
    — Amazon's new customer (Whole Foods Acquisition) https://stratechery.com/2017/amazons-new-customer/

    👋CONTACT

    Any feedback? Leave a comment or hit us on twitter
    @tom_k_white, @elliot_c_smith

    #podcast #tech #ai #startup

    56 min
  • Do we need AI in Marketing and Agriculture? | EP016

    The cursed episode… This is V3 of this episode after a couple of tech failures…

    Tom and Elliot talk about two papers, one looking at AI in marketing and another in yield prediction for grapes. Both toe the line of "did we really need AI for this".

    ⌛Timestamps

    00:00 - Start

    05:06 - AI For Marketing Misses the Mark
    40:42 - AI For Grape Yield

    ⚡Links

    — Predicting consumers’ choices in the age of the internet, AI, and almost perfect tracking: Some things change, the key challenges do not https://myscp.onlinelibrary.wiley.com/doi/abs/10.1002/arcp.1068
    — Marketers can’t predict what you’ll buy - https://fortune.com/2021/01/02/digital-marketing-ai-artificial-intelligence-netflix-amazon-limits/
    — End-to-end deep learning for directly estimating grape yield from ground-based imagery https://arxiv.org/abs/2208.02394

    👋CONTACT

    Any feedback? Leave a comment or hit us on twitter
    @tom_k_white, @elliot_c_smith

    #podcast #tech #ai #startup

    58 min
  • DALLE2 Pricing Model and Human Induced AI Bias in Language Models | EP015

    Tom and Elliot talk about DALLE-2’s new pricing model, and explore the topic of not only introducing data biases but also reasoning biases.

    ⌛Timestamps

    00:00 - Start

    02:10 - Topic 1 - DALLE2 Cost and Bias
    28:31 - Topic 2 - Human Bias translates to AI Bias

    ⚡Links

    — DALLE-2 Pricing https://openai.com/api/pricing/
    — AI Shows human like content-effects on reasoning https://arxiv.org/pdf/2207.07051.pdf 

    👋CONTACT

    Any feedback? Leave a comment or hit us on twitter
    @tom_k_white, @elliot_c_smith

    #podcast #tech #ai #startup


    58 min
  • Steve Jobs on Intelligence and AI Explainability | EP014

    Tom and Elliot talk about a Steve Jobs speech and his thoughts on intelligence and whether explainability will save us from stochastic parrots.

    ⌛Timestamps

    00:00:00 - Start
    00:06:00 - Topic 1 - Steve Jobs on Intelligence
    00:36:24 - Topic 2 - AI explainability

    ⚡Links

    — Steve Jobs - Speech to the Academy of Achievement June 1982 https://www.youtube.com/watch?v=ymbD_a-G1IQ
    — Duncker’s Candle Problem https://en.wikipedia.org/wiki/Candle_problem
    — Bullet Holes & Biases Abraham Wald https://mcdreeamiemusings.com/blog/2019/4/1/survivorship-bias-how-lessons-from-world-war-two-affect-clinical-research-today
    — Nautilus Article https://nautil.us/moving-beyond-mimicry-in-artificial-intelligence-21015/
    — Article from Lauren Oakden-Rayner https://laurenoakdenrayner.com/2018/06/05/explain-yourself-machine-producing-simple-text-descriptions-for-ai-interpretability/
    — Associated paper https://arxiv.org/abs/1806.00340
    — Moonwalking with Einstein https://en.wikipedia.org/wiki/Moonwalking_with_Einstein
    — Paper on synthetic MRI data https://link.springer.com/chapter/10.1007/978-3-030-87592-3_1
    — Paper on comparing Saliency Methods https://ieeexplore.ieee.org/abstract/document/9647419
    — BLEU https://machinelearningmastery.com/calculate-bleu-score-for-text-python/

    👋CONTACT Any feedback?

    Leave a comment or hit us on Twitter @tom_k_white, @elliot_c_smith 

    #podcast #tech #ai #startup

    1 hr 10 min
  • Legalities of Copilot & $1 trillion on health admin | EP013

    This week Tom and Elliot dig into the new Copilot feature from GitHub and discuss a recent report from McKInsey that shows how the US spends nearly $1 Trillion on admin in the healthcare sector.

    00:00:00 - Start

    00:02:13 - Topic 1 - Github Copilot

    00:03:37 - What is Copilot

    00:07:01 - What's causing a push back

    00:09:25 - What is Microsoft's plan here

    00:11:16 - Does Copilot copy paste code?

    00:12:35 - The risk for enterprise dev teams

    00:15:47 - How do I know if my code was used to train

    00:16:56 - Similarity to Oracle vs Google

    00:18:51 - Copilot protection lessons from Nintendo

    00:31:15 - Topic 2 McKinsey Healthcare Report

    00:32:47 - How we spend 1T a year in health admin

    00:36:00 - Calculating health admin spend

    00:41:05 - How'd it get so bad?

    00:46:26 - Criticisms of the report

    01:07:22 - Wrapping Up

    ⚡Links mentioned

    - Mckinsey Report https://www.mckinsey.com/industries/healthcare-systems-and-services/our-insights/claiming-the-1-trillion-prize-in-us-health-care

    👋CONTACT

    Any feedback? Leave a comment or hit us on Twitter

    @tom_k_white, @elliot_c_smith

    #podcast #tech #ai #startup

    1 hr 10 min

About High Output AI

From the publisher's feed

A show dedicated to the latest news in the world of AI & Machine Learning.