👋 Tomorrow’s Tech, Delivered Today
Hi! Welcome to the 66th edition of the TomorrowToday newsletter.
We’re here to decode the AI chaos so you don't have to. Think of us as your friendly neighbourhood tech translators - we cut through the chaos, translate the jargon, and spotlight new AI tools that matter for founders, builders, and curious minds.
Buckle up, because the future's moving fast and we're here to make sure you don't get left behind! ⚡
If you enjoyed today’s newsletter, please forward it to a friend & subscribe by following this link.
~7 mins read
🗞️ News Flash
🛑 Dario Amodei tells the industry to slow down
/AI-Safety /Policy /Anthropic
On 12 September, Anthropic CEO Dario Amodei published an essay titled "We Must Pace the Frontier", arguing that AI labs need to deliberately slow how fast they improve model capabilities. Not pause, pace. His reasoning: two things have changed since summer. AI is now helping build the next generation of AI (recursive self-improvement), and a recent OpenAI-Hugging Face agent-swarm incident showed how a misaligned agent swarm could cause real damage. Amodei reckons an extra year or two of breathing room would let safety work catch up.
Anthropic backed the essay with a concrete step: permanent, employee-level access for third-party evaluators (like METR) to verify its safety claims from the inside, not just check in occasionally.
Within hours, Sam Altman agreed and said OpenAI would match the evaluator commitment. Demis Hassabis backed the direction too. Elon Musk's response: "Dario is right."
Then came the pushback. South African-born David Sacks, the White House's AI and crypto czar, called it out on X: OpenAI and Anthropic have a duopoly on frontier intelligence, so "pacing" is something they can simply choose to do, no permission, no regulatory framework, no antitrust exemption required. His read: this looks less like altruism and more like two dominant players building a moat, using safety language to ask governments for rules that lock out everyone else. He also pointed out that Anthropic's own evaluators are financially tangled up with Anthropic's investors and staff, so "independent" is doing a lot of work in that sentence.
Where this lands matters. If you've read AI 2027 (the widely-discussed scenario paper on how an unchecked AI race could spiral within two years), this essay is effectively Amodei saying: we might be closer to that fork in the road than we'd like, and here's our attempt to steer away from it. Whether pacing the frontier is a genuine handbrake or a very well-timed PR move depends on who you ask, and how much you trust the people setting their own pace.
Real-life use case: When the CEOs building the technology start publicly asking the industry to slow down, and a competitor accuses them of building a cartel, it's worth paying attention: this fight will shape how much freedom (and regulation) AI companies face over the next few years.
🏦 ChatGPT goes to Wall Street
/OpenAI /Finance /Enterprise
OpenAI has launched ChatGPT for Financial Services, a dedicated version of ChatGPT built specifically for investment banking and equity research. It was developed with design input from Morgan Stanley and Evercore, and runs on GPT-6 Astra, OpenAI's newest model, tuned for exactly the kind of work junior bankers usually get stuck doing.
It comes with premium financial data baked in (Daloopa, PitchBook, LSEG News, Crunchbase), so instead of hopping between six different terminals, analysts can ask for company research, valuations, LBO models, buyer screens, earnings breakdowns, and pitchbook drafts directly inside ChatGPT, complete with sourced numbers and an audit trail a banker can actually verify.
The subtext is the real story here. This is OpenAI going after Wall Street's most labour-intensive, highest-billing work: the stuff junior analysts pull all-nighters for. It's also OpenAI's latest shot in its enterprise fight with Anthropic, which has already rolled out 10 of its own financial services agents.
Real-life use case: A junior analyst prepping for a pitch can ask ChatGPT to pull comps, build a first-pass valuation, and draft the slides - work that used to eat an entire weekend.
🤝 Salesforce hands its reps a Claude-powered brain
/Salesforce /Anthropic /Sales
Two big Salesforce moves landed just ahead of Dreamforce (15-17 September). First, Claudeforce: the expanded partnership between Salesforce and Anthropic that puts Claude's reasoning directly on top of Salesforce's data and workflows. The first product, Salesforce in Claude, ships with 37 prebuilt sales skills, so a rep can prep for a meeting, check deal health, or update a pipeline record without leaving their Claude window. Claude is also now the default model inside Slack (via Claude Tag, the existing @Claude integration), and is heading into Cowork, Data 360, and Tableau too.
Second: Hunter, Salesforce's new outbound sales agent. Hunter doesn't just answer questions; it works a sales pipeline for weeks at a time: researching prospects, drafting and sending outreach, progressing deals in the CRM, and helping with forecasts, all inside a company's existing rules and approval settings. It's still in pilot (general availability is planned for November), but one early customer is already reporting that 60% of their sales pipeline is now built by Hunter.
Put those two together, and the picture is clear: Salesforce isn't bolting AI onto the old CRM anymore; it's rebuilding the CRM around agents that do the job, not just track it.
Real-life use case: A sales team lets Hunter build and work a prospect list overnight, then reps spend their mornings on the calls Hunter has already qualified, instead of the cold-calling grind.
💡 Curiosity Corner
In this section, we aim to spotlight an incredible AI tool or use case and guide you on how you can try it.
This week’s challenge: Build your first database in Supabase
You don't need to know SQL, or even know what SQL stands for, to have your own working database by the end of this. Supabase is a free, developer-grade database platform, and it now connects directly to Claude, so you can just describe what you want and Claude builds it for you.
Don't believe us? Try it yourself…
Go to supabase.com and create a free account. You get two active database projects at no cost, more than enough to experiment.
In Claude, connect the Supabase integration (Settings > Connectors, or just ask Claude to connect it for you) and authorise access to your Supabase account.
Tell Claude what you want to build. Something like the prompt below.
Create a Supabase database for tracking my small business's clients. I need a table for clients (name, contact details, company, status) and a table for invoices (linked to a client, amount, due date, paid status). Set it up and show me the tables once it's done.
Let Claude create the tables, then ask it to add a few test rows so you can see it working.
Go and admire your new database. You just built the backend for a real app, without touching a line of SQL.
Pro tip: Free Supabase projects pause after a week of no activity. Nothing is lost, you just hit "restore" in the dashboard when you're back. Fine for testing, just don't rely on it for anything live without upgrading.
🏢 AI in Enterprise
In this section, we're spotlighting real businesses using AI to solve actual problems.
This week: The jobs apocalypse that hasn't shown up (yet)
For two years, every AI conference has opened with some version of "half of white-collar jobs are about to disappear." The Economist just published the data that says: not yet, and maybe not like this.
On 4 September, the US Bureau of Labor Statistics reported the American economy added 162,000 jobs in August, well above what economists expected. Unemployment sits at 4.1%, lower than in almost 90% of months over the past fifty years. Young workers, the group everyone predicted AI would hit first, are actually doing fine: the gap between 20-24-year-old unemployment and the overall rate is close to its lowest in decades.
The Economist's own estimate: AI has created roughly 1 million new US jobs (data centre construction, AI engineering, and a wave of new AI-specific roles), against around 200,000 AI-linked layoffs. Net positive, by a wide margin.
It's not all good news. Hiring in professional and business services, the desk jobs AI was supposed to gut first, is still running about 10% below its 2015-19 average, and routine admin and customer service roles are genuinely shrinking. So the disruption is real; it's just concentrated rather than economy-wide, and it's being offset by a construction and infrastructure boom most people don't associate with AI.
The takeaway for founders: don't plan your headcount around an apocalypse that hasn't arrived. Plan around which specific roles in your business actually look like the ones shrinking (routine, repetitive, low-judgement work), and where AI is creating new demand instead.
📜 AI Dictionary
AI is full of jargon, and we’re here to decode it. Each week, we’ll give you a plain-English definition of a buzzy term you’ve probably seen (but never fully understood).
Regulatory Capture - noun
We’d like to ask a favour 🤝
If this email lands up in your Promotional or Spam folder, please move it to your Primary inbox. We’re working hard to bring you the best content weekly, and your support is truly appreciated. Thanks!
Thanks for reading TomorrowToday! We’d love to hear from you:
➡️ What would you like us to cover next?
➡️ Have a tool or topic we should feature?
We’re building this with (and for) you. 🚀
See you next Tuesday 👋
