Tom Siwik

I don't care if you have 3 followers or 200k, I will follow you because your feed seems interesting to me

I don't understand why people are so hostile against technologies that solve problems just because it originated from a big corp. Graphql is bad, React is bad, MCP is bad, Java is bad... it's exhausting. If you don't like an aspect of it, solve it and mold it to your standard

@kimmonismus Companies are cashing out while firing. Not firing because of AI. They are using this narrative beneficially. Big difference

Anthropic co-founder Dario Amodei has been saying this for over a year now. And he keeps saying it. Louder each time. In May 2025, he told Axios that AI could eliminate 50% of all entry-level white-collar jobs within five years and push unemployment to 10-20%. In January 2026, he published a 20,000-word essay calling AI “a general labor substitute for humans” that will cause “unusually painful” disruption. At Davos, he warned of a “zeroth world country” forming in Silicon Valley, decoupled from the rest of society, running at 50% GDP growth while everyone else faces mass joblessness. In his own words: “We, as the producers of this technology, have a duty and an obligation to be honest about what is coming.” And the data is starting to back him up. Tech entry-level hiring dropped 30-50% in 2025. Wall Street banks are cutting ~200,000 roles concentrated at the junior level. S&P 500 companies shed employees in net terms for the first time since 2016. Anthropic’s own labor market research confirmed that 77% of businesses use Claude to automate tasks, not to augment workers. Now another Anthropic co-founder is echoing the same message: “There is a real possibility that AI will displace human labor at a very large scale. Supporting those people will be a moral imperative of historic proportions.” This is no longer a warning from the sidelines. This is the company building the technology telling you, repeatedly, that the disruption is real, it’s fast, and society is not ready for it. x.com/disclosetv/sta…

@theo You can hide this quirk gradually increasing the timeout and still be unhackable

Because if it’s fast, I can spam until I get the right password and hack you. This is a good thing.

@izadoesdev What's missing?

AI SDK has kinda fallen off, no more cool releases or updates, no innovation, commits slowed down, what's up with that?

@stemonteduro tailscale, slightly less & more secure setup than cf tunnels

If you still think you need to sit at your desk to build stuff, you’re ngmi. I’ve never seriously experienced a full shipping cycle from mobile before. And it’s amazing. I came up with the idea right before leaving for an event in Otranto. So I opened the remote Claude session and just kept chatting and building from there. I also have Cloudflare Tunnels pointing to my local project, so I can basically test everything remotely from my phone. What I loved the most is that I could just pull my iPhone out of my pocket whenever I had a free moment and continue building. And honestly, it’s amazing because when I’m working from my Mac, there’s so much dead time while the agent is cooking anyway. I might actually get addicted to this.

@DmytroKrasun If your api doesn't manage a lot of cross-entity dependencies it works out fine for most cases. glad it worked out too

JSON:API is beautiful. I once shipped and worked with it on production in a very complex system. But in my case, I didn’t expect my API to survive for so long 😆 So versions didn’t make sense either. Plus, versioning is always hard. However, I intentionally tried to keep everything simple and backward compatible and that worked out well.

I grew ScreenshotOne to over 900 paying customers in more than 3 years with 2 API endpoints and around 30 options. For an API product, adding new features is a huge problem. Every new endpoint or option you add will stay with you forever. Forever! Nobody will migrate and spend time on your breaking changes. And you don’t want to be that annoying API provider that breaks something every week, and everybody complains about. You want your product to be synonymous with reliability. Damn, I don’t even have API versions. I support every option I have ever added and keep them all backward compatible. I am heavily inspired by Java and Go. Old code should keep working. But there is a price for all that. It is unimaginably hard to implement feature requests. And yet you have large paying customers who ask for more and more. My current approach is: - have a vision of where I am going; - add primitives slowly; - make them reusable for future capabilities. For example, many customers asked to add full-page screenshot slicing. I didn't want to add it and make the API more complex. But now when I have a vision, the feature perfectly fits it. ScreenshotOne is becoming a reliable vision layer for agents: give a URL and get what you need from it. The same feature will later power more agentic workflows, but it will be battle-tested in production first. Build simple primitives aligned with your vision that, in combination, make customers powerful and help them extract more value from your product.

Painful experience where google is regex-filtering out links (it's own too)

@rcwhiteley I think there is a reason animals identify each other by smell instead

Facial recognition for animals. Who's building this? It should be no more difficult than for people. Could even do it for plants, identifying individual organisms.

@DeniCodes If you want them to output in a specific way and not ramble in complex jargon. E.g. adversarial review, idea generation, deep researcher.

@tomhacks Good to know what would be some use cases for non normies for this?

saw a few younger guys on the metro talking to chat gpt, the first sentence was “you’re an expert manager…” people still prompt like “you’re an ExpErT software engineer” stop, get some help, this isn’t 2022 anymore Just type whatever you need and get on with it, LLMs know

@woocassh Yepp last visit already. Delish

@tomhacks Heeey Tom on holidays?! Try pastel del nata yet?

Found the biggest Asparagus in Nazare/Portugal If you eat this one, your pee will smell until your end of life

@AlxTurovski Meh. Enough for surfing

@tomhacks Lol How’s the weather btw?

@unclebobmartin how much effort do you put in either side? more on the test-side or the implementation? I'd love to truely let go of one side over the other but the quality suffers when I start vibe coding. So I HAVE to handhold the AI constantly which does not feel productive at all.

I am absolutely more productive using agents. I don't know the factor but it's large. However much of that productivity is spent tuning the agents and hardening the product. I'm guessing 30%-40%. Some might consider that a waste; but I don't. The software I'm creating nowadays is vastly more robust than I'd ever been able to create manually. I don't mean that the code is better. I mean the surrounding tests are vastly better. I have a higher degree of confidence than I ever had manually -- even when I used very disciplined TDD and Acceptance testing. And then there's the ability to quickly reorganize the modules and the architecture while keeping those robust tests running. That is a tremendous boon.

4m height btw.

@tomhacks tom's followers: 💹 tom's reach: 📉 tom's content: 🥇

Why did I get this now? Every blue checkmark is fake. but it buys you reply boosts and ppl are bad at self-psychology

@theo @umohsg Gov + Cookie Consent

Crazy how the web is still the only platform that doesn’t suck to develop for.

@ThePrimeagen At this rate limiting you can't even go for 2x engineer

Honestly why stop at 100x engineer? Just use more agents, you literally could be 1000x, 10000x, 100000x just by scaling You could what you use to in an entire year in one second

@bananawalnutz something like that. yes

@tomhacks interesting, so like samsara in all possible incarnations until “God Consciousness”?

can we just accept that consciousness is recursion and move on?