Gijs

363 posts

Gijs

@datagobes

Senior cloud data & AI engineer. Building https://t.co/SGOypsko6x in public — AI-native tools, MDX blogs, and things that probably shouldn't work but do.

Amsterdam Entrou em Mart 2026

128 Seguindo76 Seguidores

Tweet fixado

Gijs@datagobes·29 Mar

My free AI jukebox has been running for 2 weeks now. 50 community generated tracks, hundreds of plays. I’ve added audio visualisations, full screen mode and a docked player that works across my website. Come take a look and generate something weird! datagobes.dev/community/juke…

English

1.3K

Gijs@datagobes·4m

@BacLeodiv Did you look into the new effort settings? It no longer defaults to medium. Apparently there is also a new tokeniser in play that can increase the number of tokens up to 1.35x for the same text.

English

Bac Leo@BacLeodiv·1h

After nearly half a day of testing, I believe: Opus 4.7 = Opus 4.6 + 50% more expensive Is it just me, or does anyone else feel the same?

English

190

Gijs@datagobes·13m

First you need to find something that’s your passion or you will not make it to consistency. I had to play catch-up in my thirties because I didn’t leverage my twenties properly. Tried to do things because others were doing it and computer stuff was for nerds. I went to army officer academy, studied international business, only to come back to my true passion: computer nerding.

English

andrei saioc@asaio87·1h

If you are in your 20s, then in this fast-paced AI world, you need one single thing for the next 5 years: CONSISTENCY stay consistent for the next 5 years with whatever you are doing and keep improving. I would do this: - create content online about what you do - do it every single day - do short form and long form now with AI is super easy to get some ideas, create a content plan for the whole month. yes it might take you 30-60 mins to produce one 10 min video a day yes the first 100 videos will absolutely such but you will learn so much, and you will become 100x better after the first year

English

Gijs@datagobes·18m

@KaiXCreator Claude for the overall win

English

Kaito@KaiXCreator·1h

Which one is the worthy investment?

English

487

Gijs@datagobes·47m

@GamewithDave James Pond 2

English

Dave@GamewithDave·16h

For anyone who used a computer between 1990 & 2005… what’s the one game you still think about?

English

36K

536

11.6K

Gijs@datagobes·1h

@trikcode Manual coding going away doesn’t mean you can have monkeys build your software. You still need engineers, capable ones.

English

Wise@trikcode·1h

Anthropic's CEO said coding is going away. Anthropic is currently hiring 454 engineers at $320K-$405K. who's writing the job descriptions

English

147

6.2K

Gijs@datagobes·1h

In the new Claude desktop app I have a bit more success with it due to how it organises and flags the threads you have running. But overall I agree. I went down this rabbit hole, chasing some illusion of ultimate parallelism but in the it just burned me out and I had to take a break.

English

andrei saioc@asaio87·1h

I think 99% of people orchestrating AI agents are either lying about their success rate or are really burnt out. Its quite hard to focus your attention on 2-5- 10 different places in one go, because these agents need to be babysat When you add everything up, you are indeed gaining more time but its not at a 5-10x order of magnitude. I would say all in all might be 10-20% time gain.

English

402

Gijs@datagobes·9h

@kaaaash____ Yes, since Claude I push code to GitHub again. It has been years..

English

Akash@kaaaash____·22h

Do you still push code to GitHub after your job?

English

339

Gijs@datagobes·10h

@Shubhyxz A colleague of mine always called me with pyspark questions. At one point he asked: how do you know this? Me: I just googled it The colleague: Ah, so you’re not good at pyspark at all, you’re just good at googling. 🤷‍♂️

English

Shubh@Shubhyxz·18h

90% coding 10% Googling 100% pretending I knew the solution

English

Gijs@datagobes·10h

@0xPrajwal_ @md_kasif_uddin Here’s a sharepoint site bro. GLHF

English

Prajwal@0xPrajwal_·17h

Interviewer: Do you know system design ? Candidate: Yes! Interviewer: Design a system for 100 users... Candidate: Microservices, load balancer, queues… Interviewer: You are solving for millions... I only asked for 100.

English

222

33.5K

Gijs@datagobes·10h

Vibe coding makes you fast at building things you don’t understand. Most of you will ship for 6 months, hit the first real problem, and realize you have no idea what you actually built. The tool isn’t the bottleneck. You are.

English

Gijs@datagobes·11h

@dramaricic I have a different experience. I’ve already made 5 million on 4.7. Follow me to learn how.

English

Dragan Maricic@dramaricic·13h

So Opus 4.7 is just normal Opus 4.6…

English

1.7K

Gijs@datagobes·11h

I’ve just sent emails to 10 big companies containing a summary of the findings of my gdpr privacy scanner. 1 company already responded within 24 hours they want to see the full report. In two weeks, when all companies will have had ample time to respond to my findings, I will start publishing the scan results. It’s a big mess in privacy land. Read the blog: datagobes.dev/blog/privacy-a…

English

Gijs@datagobes·12h

@asaio87 @schedpilot Cool project! Keep it up

English

andrei saioc@asaio87·21h

Most founders are too busy to be consistent on social media. You need to keep posting every single day and on multiple platforms. Even one platform is a lot. I build @schedpilot to keep up

English

371

Gijs@datagobes·15h

@KaiXCreator You: Python is not used to run the inference runtime itself. They only give it a python API so it can easily be integrated in the wider python ecosystem.

English

Kaito@KaiXCreator·16h

Interviewer: Python is slow compared to C++. Then why do top AI companies use it to build the world's most powerful models?

English

1.8K

Gijs@datagobes·1d

If you don’t want to get copied over a weekend here’s my two cents. Don’t go into a craft where you don’t have any expertise just because AI lets you generate plausible results. Figure out how to use AI in the areas where you already have expertise. Then partner with people if you are missing expertise somewhere. Don’t use AI to amplify your weaknesses, use it to amplify your strengths.

English

Gijs@datagobes·1d

@thedevchandra Liked, followed. Here's what I am working on: document intelligence. I am a data engineer originally and now also applying AI in a corporate environment and mentoring other business roles in applying AI. x.com/datagobes/stat…

Gijs@datagobes

Andrej is right. Processing pdf's is hard but lots of knowledge is captured in them. I am focusing on using AI in a corporate setting and document intelligence is a challenge. I ran into this problem when I first started designing an agent that could design agentic systems. I wanted to put all the big papers in a knowledge base and it didn't work out. Until I created Felix, my document intelligence project. Felix decomposes any business document into typed elements, stores them in postgres, and can reconstruct the full document from it. Now that Andrej dropped his llm-wiki idea, I've added an mcp on top so I can use Felix for llm-wiki generation and I will post my wiki from the Mythos document on GitHub later on. (Claude happened to go down the moment I was generating it) See the screenshots for a UI view on the Felix API, which is showing the document Andrej is mentioning here, decomposed and served from postgres. I am not going to tell you to reply FELIX and follow me to get some nonsense prompt and a 199 offer for a crappy workshop. But I am not going to stop you from letting me know you think this is interesting either. If this gets some engagement I will package it and put it on GitHub.

English

Dev@thedevchandra·1d

I'm tired of seeing the same faces on my feed man I need brothers in tech, AI, startups, marketing, distribution, vibecoding, sf or elsewhere to build along like, drop your startup, let's make some friends.

English

424

721

26.3K

Gijs@datagobes·1d

@cet3001 @pmitu Exactly this. You need to keep improving the context until you get consistently great output.

English

Calvin Thurman@cet3001·1d

@pmitu I'd push back on this slightly. Some people do need AI to think FOR them first just to get started, then they refine it. The problem isn't people using AI to think. The problem is people stopping there and treating the first output as the final answer.

English

Paul Mit@pmitu·1d

Don't let AI do the thinking for you! Use AI to think more deeply, harder, and more accurately.

English

125

2.4K

Gijs@datagobes·1d

@asaio87 For reference, this was my result. This is the light theme it did and it has a dark variant too. I can see now your ember is a bit more towards red but the background, its hard to tell if its not the same.

English

andrei saioc@asaio87·2d

@datagobes haha no, it was a 2 sentence prompt. The old one looked awful so im happy with this, just to have info put out there in a nicer way

English

andrei saioc@asaio87·2d

I just vibecoded my personal website. Well i redone it, it was looking awful because it was vibecoded with a free version of chat gpt a year ago. Now i made it better with the paid version of Claude. I am happy enough to just have it out there with all my information, all my accounts and some info about myself. Ah and I have a newsletter too Link in the first reply

English

501

Gijs@datagobes·1d

@gregisenberg Might work for a few. I am pretty convinced right now it's best to stick to your expertise. If you need software, find somebody who can already do this. then combine both your expertises to create something unique. This way you get good software that also solves a real problem.

English

159

GREG ISENBERG@gregisenberg·1d

agents are the new apps the dirty secret of the SaaS era is that the software never actually worked. it was always 70% product, 30% the specific person in your company who knew how to make it behave. that person was called a "power user" they were actually just a human patch agents replace the patch and suddenly everyone realizes the software was broken the whole time what's cool is how much opportunity there is right now so pick a niche. any niche that you believe 1% of the market is $5M ARR+ there the leader in that space has a 20 year old codebase and a customer base that only stayed because switching was painful and that pain just got a lot easier to swallow agents are the new apps are you building yet

English

648

35.2K

Gijs@datagobes·1d

@Gamingtronium @misschopra Gardening and plumbing is where it’s at right now