The Ledger
Every statement that passed quotation and attribution checks. Mix any filter with any other: certainty 1/5, debate potential 5/5, or both at once.
why aren't all 99 resolved? a statement only gets an assessment when the public record can support or contradict it. opinions and what-ifs never can, and 0 checkable ones are still open, waiting for their date. predictions held up or didn't; assertions are supported or contradicted. on every card: ▮▮▮▮▮ certainty · ▮▮▮▮▮ debate potential. speakers are clickable
Tom Blomfield says today's software engineering jobs will vanish within 10 years
“But the conclusion of all of that, I think, is that software engineering jobs of today, I think, will not exist in five or 10 years.”
Tom Blomfield says AI will fulfill surging global demand for software
“Like I could easily see the demand goes up 10 X or a hundred X, who knows, but I just think the AI is getting so good that you just won't have humans fulfilling that demand in anything but the most niche of niche cases.”
Blomfield: Rebranding Twitter to X was total insanity
“I think when you get something short and catchy, and you have the verb, just like total insanity to change, especially for something as meaningless as X, right?”
Blomfield: Discussing pricing early almost never scares off enterprise prospects
“Founders are often very afraid to have a willingness to pay conversation. They feel like if they bring a dollar amount up, it might scare the customer off. But honestly, that's almost always not true.”
Blomfield: Early startups should accept customer contract terms unless company-ending
“So I'd be pretty flexible on signing, frankly, whatever your early customers want, as long as it's not going to expose you to like unlimited liability or container clause that's going to transfer all the IP in your product to your customer or something company…”
Blomfield: Embedding chatbots into software is almost always the wrong approach
“I kind of hate the kind of chatbot paradigm. Like it was the easiest It was the way to bring the most basic LLM experience to the whole world. You know, we had GPT before ChatGPT and ChatGPT brought it to the mainstream, but now every product is just embedding…”
Tom Blomfield predicts stark consequences for software engineering within 12 months
“I think we have a clear line of sight between where we are today and like improvements over the next three, six and 12 months. They're going to have quite stark consequences for software engineering, I think.”
Tom Blomfield built a 35,000-line app without writing any code manually
“Yeah, 35,000 lines, and I wrote not a single one. And honestly, after about the first 5000 lines, I stopped even reading the code.”
Tom Blomfield says future engineers will primarily focus on wrangling AI machines
“I think there will be demand for smart people who know how to wrangle these AI coding machines, and if we want to call those people software engineers, so be it. But I think the job is dramatically, dramatically different.”
Tom Blomfield says software startups are reaching revenue milestones faster than ever
“So getting to a 1,000,010 million, a hundred million of revenue with a software idea had is just happening faster and faster than ever.”
Tom Blomfield warns AI could displace hundreds of millions of workers
“And you could see hundreds of millions of people displaced, and the idea that they're going to retrain as in a different job, I think is going to be extremely painful. And I can see the societal impact and turmoil being very, very grave for 10 or 20 years as t…”
Blomfield: A +50 NPS is the minimum baseline for new consumer companies
“If your net promoter score, your MPS, is not extremely high as a new consumer company, you're toast. You pretty much have to re-engineer the product to make it something that people love. The reason is it correlates extremely well with word of mouth referral. …”
Blomfield: Google lost its transformer lead to OpenAI through excessive caution
“And so, I think, like, Google were really on the cutting edge of this transformer technology, and were like, leapfrogged by OpenAI, just because I was too cautious to put stuff out there.”
Blomfield: In 5 years, 99% of people won't touch their system prompts
“But I, I'd suggest, I think in In five years, for 99% of people, they're not touching their system prompt, but the system prompt is custom to them.”
Tom Blomfield says AI makes him 10x more productive than before
“These new tools made me 10 times more powerful, more productive than I was when I was kind of current in, in, you know, software engineering 10 years ago.”
Tom Blomfield says AI coding in YC startups surged to 25%
“Last batch, that number was 25% of companies. So they use these tools for most of their code. Two batches ago, it was approximately zero percent.”
Tom Blomfield predicts trade bodies will act as protectionists against AI
“And so I can see that happening in, in law, in medicine, as these kind of trade bodies act as gatekeepers to try and, almost like protectionist unions, trying to protect the jobs of their members.”
Tom Blomfield says AI will transform law and medicine within five years
“So many industries like law or education or medicine that really were not big buyers of software in the past that are going to be transformed in the next five years.”
Blomfield: Lovable struggles when modifying backend logic rather than UI
“Tools like Lovable started to struggle when I wanted to more precisely modify backend logic. Rather than just pure UI changes. I'd change a button over here, and the backend logic would bizarrely change.”
Blomfield: Using MCP servers for API documentation is overkill
“Using an MCP server to access this documentation, which works for some people, seems like overkill to me.”
Blomfield: AI Coding Might Shift Software Architecture Toward Modular Services
“I think we might see a shift towards more modular or service-based architecture where the LLM has clear API boundaries that it can work within while maintaining a consistent external interface Rather than these huge monorepos with massive interdependencies.”
Blomfield: Gemini Leads Codebase Planning, Claude 3.7 Leads Code Implementation
“For example, at the moment, Gemini seems best for whole code-based indexing and coming up with implementation plan, while Sonnet 3.7, to me at least, seems like the leading contender to actually implement the code changes.”
Blomfield: GPT-4 Underperformed on Coding by Misimplementing and Asking Too Many Questions
“I tried GPT-IV just a couple of days ago, and honestly, I wasn't yet as impressed. It just came back with me with too many questions and actually got the implementation wrong too many times.”
Blomfield: Community Notes has done a good job flagging fake news
“And I feel like community notes have done a relatively good job of calling out the stuff that's fake news versus real news actually. So in, in one sense, I think that's been a really positive change.”
Blomfield: Committed recurring revenue is preferable to usage-based pricing
“In general, committed recurring revenue, that's monthly recurring revenue or annual recurring revenue is preferable to usage based pricing. And that's because during an economic downturn or a slowdown, your revenue is protected at least until the contract is u…”
Blomfield: Complex multi-touch attribution is more hassle than it is worth for startups
“You can get really convoluted with, you know, first touch attribution, last touch, multi-touch attribution. It's often, ah, way more hassle than it's worth. Simply using those UTM referrers or asking your customers where they came, where they heard about your …”
Monzo scaled to 500k users with unit economics of -£30 to -£40
“In the early days of Monzo, we were at -30 or -40 pounds per customer per year, and we scaled to half a million customers before we fixed them.”
Blomfield: AI startups relying on OpenAI credits face severe margin shocks
“Just because you're getting free credits doesn't mean that that's a cost that doesn't exist. It just means you're hiding it for the moment. So companies that hide behind OpenAI credits and claim that they've got these huge, huge gross margins have a nasty shoc…”
Blomfield: Monzo lost £30–£40 per user initially before turning profitable
“Monzo is an online bank in the UK, and for our first half a million customers or so, we were losing money on every customer, 30 or 40 pounds per customer. We scaled to more than half a million of those customers. It costs a lot, but we had a plan to turn it ro…”
Blomfield: Founders must fix negative unit economics before scaling
“So if you start with negative unit economics, you really, really have to have a plan to fix them. And I would really advise you don't scale your customer base. You don't try and grow as quickly as possible whilst you have negative unit economics. You fix them …”
Blomfield: 90% of B2B Founders Get Stuck in Unpaid Design Partnerships
“And really, 90% of the time, the vast, vast majority of the time, founders get stuck in the very early stages. It's the very long unpaid design partnerships.”
Blomfield: Early startups waste time overbuilding platforms instead of narrow wedges
“What many founders do, which is a mistake, is to try and overbuild a really broad platform, which is a mistake at this stage of your company because you just don't have the resources. You can waste a lot of time without any real signal that a customer wants wh…”
Blomfield: Free B2B trials fail when run without clear end goals
“The most common problem for these free trials is again, they're too long, maybe two or three months, and they suffer from the same low commitment problems as a design partnership. There's no target or end goal, and the customer isn't really committed to engagi…”
Blomfield: Early startup sales sell the founders, not a bug-free product
“Because your product is so early, you're never really selling a complete bug free experience. You're actually selling the founders and the early team. You're selling the promise that you personally are going to solve this problem for them.”
Blomfield: Reducing time-to-first-value is the top lever for pilot conversion
“Reducing it from weeks to hours is often the biggest single lever for higher pilot to paid conversion.”
Blomfield: Never require customer engineering time during a sales pilot
“I would never do a full API integration during a pilot if it requires any engineering time from the customer whatsoever, because that's just going to delay you by months.”
Blomfield: Use Recurring Contracts With Opt-Outs Instead of Standalone Pilots
“So to get around this really sophisticated founders move on from paid pilots to recurring revenue contracts with an opt out period. So this is typically a monthly or annually recurring contract With a 30 or 60 day money back guarantee or opt out period at the …”
Blomfield: Coding agents are far ahead of other general-purpose AI agents
“The coding agents especially just seems so far ahead of all the other general purpose agents right now.”
Blomfield: Training AI assistants requires gradual feedback, not 30-page upfront instructions
“The idea that, take a human analogy, you hire a new employee, you hire an assistant, the idea that you're going to sit down and write out 30 pages of instructions of how you do every single thing in your life is possible, but none of us actually do it. You sor…”
Tom Blomfield built and migrated his entire blog in 90 minutes
“In 90 minutes on a train journey with Claude Code, and in 90 minutes I set up hosting, I wrote new blogging software, and I migrated, like, 15 years of blog posts over to a new platform.”
Tom Blomfield says ignoring AI will be a disadvantage within two years
“I'm not sure it's yet a competitive disadvantage, but I can imagine that becoming true in the next year or two as your tools get really, really good.”
Tom Blomfield says early AI adopters will gain a multi-year advantage
“And if you are one of these people who are at the cutting edge of this, these tools, I think you have an advantage for several years, which will persist and enable you to frankly, to earn a lot of money to grow a great career.”
Blomfield: Product managers and designers are bypassing Figma for direct coding
“Many product managers and designers are actually going straight to implementation of a new idea in code, rather than designing mockups in something like Figma, just because it's so quick.”
Blomfield: Multi-prompting AI to fix bugs accumulates layers of bad code
“I found I had bad results if I'm, like, prompting the AI multiple times to try and get something working. It tends to accumulate layers and layers and layers of bad code rather than, like, really understanding The root cause.”
Blomfield: Git reset and re-feed final AI solutions onto clean codebases
“You might go and try four, five, six different prompts, and you finally get the solution. I'd actually just take that solution, git reset, and then feed that solution into the AI on a clean code base, so you can implement that clean solution without layers and…”
Blomfield: Claude Sonnet 3.7 configured DNS and Heroku, accelerating progress 10x
“For example, I had Claude Sonnet 3.7 configure my DNS servers, which is always a task I hated, and set up Heroku hosting via a command line tool. It was a DevOps engineer for me and accelerated my progress like 10 x.”
Blomfield: Extensive instruction files make AI coding agents far more effective
“Each tool has a slightly different naming convention, but I know founders who've written hundreds of lines of instructions for their AI coding agent, and it makes them way, way, way more effective.”
Blomfield: Storing API docs locally yields more accurate LLM code generation
“So I'll often just download all of the documentation for a given set of APIs and put them in a subdirectory of my working folder so the LLM can access them locally. And then in my instructions, I'll say, go and read the docs before you implement this thing. An…”