|
|
Hi Sparklers!
Did you see the White House lunch? Thirty-one guests, a lot of very important men, and one big promise to keep AI safe. The bosses of the six AI companies behind it are all men. Lisa Su, who runs AMD, was one of the very few women in the room. Hold that thought for story 3.
This could be one of the most important technologies of our lives, shaping our health, our work, our businesses and how we live together. It definitely needs regulating. Right now, itâs in the hands of people behaving like teenage boys left in charge of the house for the weekend.
Tonight is Switched ONline, my first free online Switched ON session, at 6pm on Zoom. Next Monday, Iâll see some of you in person in Old Street. Both are free, and both are at the bottom of this email.
First: what AI might do to our jobs, that White House promise, and a tool that turns your rambling into a full brief.
|
|
THIS WEEK
Switch ON to AI News
9 to 5
Anthropic, the company behind Claude, has built a tool asking what AI could do to US jobs by 2030. It splits a nurseâs day into tasks: some only a human can do (âAI canât bathe a patientâ), some AI helps with or takes over, and new ones, like checking how well an AI triages patients. You can plug in your own predictions. In its scenarios up to AI roughly doubling the normal rate of growth, unemployment stays within the historical range. Its small print says actual outcomes âmay differ materiallyâ.
My take: I think the Anthropic thing is very interesting. They produce these reports regularly, but I like the way theyâre trying to give people more control, more hope and real reassurance about technology. Itâs balanced. Not just doomerism, but a better sense of how things might look.
Read the full story â
Promises, Promises
On 29 September, President Trump and top AI executives signed a White House accord that leaves safety mainly in the companiesâ own hands. âIâm seeing tremendous self-policing,â said the President. The next day, Americaâs consumer watchdog, the FTC, opened a broad investigation into the safety of AI made by Anthropic and OpenAI.
My take: Theyâre marking their own homework. Pinky promises, made by the Billionaire Boys Club. Weâre cooked.
Read the full story â
Top of the World
Fei-Fei Li, one of AIâs legends, is joining AMD as executive vice president and chief scientist. The chip company is buying World Labs, the AI company she co-founded, in an all-stock deal worth around $8.2 billion, announced on 30 September. Her new boss is AMDâs chief executive, Lisa Su.
My take: Sheâs phenomenal, the Godmother of AI, and AMD has a female chief executive in Lisa Su. I think this deal is a big boost for technology, and an even bigger one for women and what might be possible. The idea of Fei-Fei and Lisa working together excites me, and I think it shows great hope for the future.
Read the full story â
Honey, I Shrunk the Free Plan
Google is limiting free Gemini to its smallest model, Flash-Lite, from Friday 9 October, according to reports. Free users lose the bigger Flash and Pro models. Google hasnât put it on its own help page yet. Most people never check which model theyâre on. From Friday, itâs worth knowing.
My take: I think the real conversation is the battleground over price and value. Itâs also a warning sign about how restricted we are in which model we use, and how vulnerable we are to price changes.
Read the full story â
Pretty Woman
ChatGPT has rolled out Try On globally, on desktop and mobile. Upload a selfie or a full-length photo, add an image of something youâd like to wear, and it shows it on you. It keeps your photo for next time, and you can add, replace or delete it in Settings, then Personalisation, then Reference photos. If you canât see Try On yet, itâs on its way.
My take: I gave it a go: from my white skirt to a new black one from Selfridges. Thatâs the AI version on the right. This kind of technology has existed elsewhere, but building it into AI is fascinating, and I think it has huge potential for shopping. It shows how serious OpenAI is about getting deeper into the way we shop.
Read the full story â
|
|
SPARK SWITCHED ON
Ramble in, brief out
Last week I gave you three lines to tack onto a messy voice note so the AI tidies it into a tighter prompt. Wispr Flow has just turned that into one button.
Itâs called Prompt Engineer. Say what you want, the messy way it comes out. Highlight it, press one shortcut (Option+2 on a Mac, unless you change it) and Flow rewrites it on the spot as a brief the AI can follow: what you want, who itâs for, the background and what you want back. It works in every AI chat app, and it appears where you type, before anything sends, so you can edit it first.
Same rambling sentence from you. The AI gets the whole assignment.
Itâs a really good build on last weekâs tip, because the newer versions of Claude and ChatGPT are rethinking the way we prompt, again. Prompt Engineer takes what would have been a really long ramble into Wispr Flow and makes it tighter, more relevant and more succinct for the AI. Probably 75% of my Wispr Flow use is prompting AI, so for me this is a big deal.
You can try Wispr Flow free for 90 days. Full disclosure: Iâm an official ambassador, and thatâs my link.
Try Wispr Flow free â
|
|
TRY IT THIS WEEK
Switch this ON
Open Claude, ChatGPT or Gemini and paste this in:
THE PROMPT · COPY IT AS IT IS
Interview me, one question at a time, until you could introduce me the way my biggest champion would. Ask about my work, my wins, the moments that changed my direction, and how I like to sound. Stop after fifteen questions. Be honest. Don't flatter me.
Answer however suits you, typed or out loud. Then save your answers in a Claude project, a ChatGPT project or a Gem in Gemini, so you donât start from scratch next time.
|
|
THE REAL TAKEAWAY
Itâs a Manâs Manâs Manâs World (For Now)
Thereâs no doubt this is one of the most powerful, seismic shifts in technology and business. If there was ever an argument for having more women involved in AI, it was the sad scenes at the White House.
Anthropicâs jobs explorer shows how AI might touch every job, and we need more and more practical examples of how AI can help everyday people. We need to stop the doom-mongering. That doesnât mean we donât accept the fears, and there are many. The doom-mongering just isnât helpful.
Tonight and next Monday are two practical places to start, and the details are just below.
|
|
OCTOBER
Come and find me
Tonight
Switched ONline
Monday 5 October · 6pm to 7.30pm UK · Live on Zoom · Free · Limited seats
You asked AI for a two-line reply and got four paragraphs, a yes to everything and the word âdelightedâ.
You are not behind. Tonight I open my laptop and you watch me hand AI the inbox and the admin. You leave with a habit that sticks, and you can ask me anything. No card, no catch, no hard sell.
From âI feel behind on AIâ to âoh, I can do thatâ.
Bring your questions. Iâll bring the sparkle.
Next Monday, in London
Wait, You Too?
Monday 12 October · Doors 6pm · Tech Junction, Old Street · Free
Women in tech? That includes all of us now.
Bring your mum, auntie, daughter, niece or friend, or come solo. Think of it as the best conversation about AI youâll have all year, with snacks. Answers to the questions youâve been too polite to ask, and nobodyâs the expert. Sequins encouraged. Allies welcome.
Come for the stories. Stay for the âwait, you too?â
Part of Women in Tech Week, supported by Civo.
|
|
HOW WAS THIS WEEK'S SPARK?
with love and sparkle,
Nish x
P.S. Switched ONline starts at 6pm tonight. Free, on Zoom, limited seats. Grab yours, then forward this to the friend who keeps saying âI feel behindâ.
|
WRITTEN BY MOUTH
My words in this newsletter started out loud, and Wispr Flow did the typing. My keyboard has never had it so easy.
Get 90 days free: wisprflow.ai/nishma. Full disclosure: I'm an official ambassador, and that's my link.
|
LinkedIn ·
Instagram ·
Substack ·
YouTube
|
|
|
