Der.dev 🔥🛠️ (@erd0xbc) header
Der.dev 🔥🛠️ (@erd0xbc) profile picture

In it for the tech

Joined June 2009
Photo by Der.dev 🔥🛠️ (@erd0xbc)Photo by Der.dev 🔥🛠️ (@erd0xbc)Photo by Der.dev 🔥🛠️ (@erd0xbc)Photo by Der.dev 🔥🛠️ (@erd0xbc)Photo by Der.dev 🔥🛠️ (@erd0xbc)Photo by Der.dev 🔥🛠️ (@erd0xbc)Photo by Der.dev 🔥🛠️ (@erd0xbc)Photo by Der.dev 🔥🛠️ (@erd0xbc)Photo by Der.dev 🔥🛠️ (@erd0xbc)Photo by Der.dev 🔥🛠️ (@erd0xbc)
Der.dev 🔥🛠️ retweeted
ok gl
2,135
1,139
447
27,482
4,379,071
Der.dev 🔥🛠️ retweeted
Tomorrow.
373
468
315
7,937
527,763
Der.dev 🔥🛠️ retweeted
past week ships: - hex, great way to talk to agent - browse, great way for agents to use your browser - rune, great way to build native apps - deepseek $100 for $10 - new opencode website unfortunately wrote zero essays about what a software factory is so they will be punished
107
58
13
2,699
87,453
Der.dev 🔥🛠️ retweeted
Replying to @JamesMalsawm
Pretty wild to me that GPT 6.1 Terra gets compared to Opus 5.5. No doubt that it's a great model, but it's not meant to compete in the big leagues. We've got other models for that.
397
34
40
2,932
289,323
Der.dev 🔥🛠️ retweeted
Most orchestration like this is for human ergonomics. More layers means more bureaucracy and far more token burn, with little to show for it. Cut middle management layers out like the plague, but do try and find systems that help you understand and review large bodies of work.
I've changed how I work over the last few weeks. I now have one lead agent for each area of focus, running all the time, and a chief of staff agent that oversees them. Each lead manages its own sub-agents that write code, test it, chase reviews, and coordinate with team members. Everything I need to decide or review comes into one prioritized queue, a simple local web app. Each product review includes screenshots and a video demo of the feature. I clear the queue in batches, mostly by voice and keyboard shortcuts. The leads also bring back ideas I hadn't considered. Working from one queue, instead of switching between dozens of chats, has cut a lot of mental overhead. The system keeps learning my preferences and distilling my judgment into a "Brian brain," so its recommendations and first drafts get better before they reach me. The goal is fewer corrections per change. Every change still goes through the same human review and security checks. It is eliminating tasks, not human judgment. What an exciting time to push the frontier of productivity, especially as we build the economy for agents (agentic finance). We should all keep pushing in this direction.
25
6
1
199
27,277
Der.dev 🔥🛠️ retweeted
Replying to @EnoReyes
i am automerging thousands of PRs already, no centralized factory needed
26
3
2
232
23,970
Der.dev 🔥🛠️ retweeted
Cyber
I guess now that @openai has name dropped @calif_io as an early adopter I can tell you GPT-6 Cyber is out of this world. It is lightyears ahead of everything I used.
425
40
22
1,690
633,315
Der.dev 🔥🛠️ retweeted
should've used pstack
1
1
19
2,231
Der.dev 🔥🛠️ retweeted
Get me out of the basement poteto.com/
1,656
222
343
8,646
2,667,240
Der.dev 🔥🛠️ retweeted
Credit where it’s due, this guy has been an industry leader in shitty takes and bait posts for almost a decade now.
Anyone else stopped vibe-coding? I've cancelled all my AI subscriptions. After initial excitement, I felt AI couldn't do more than before. It's more of a "knowledge refactor" - like an encyclopedia that reorganizes information, yet unlocks few opportunities. A dip is coming.
100
50
2
2,643
142,424
Does Europe still want to play any role in the AI race? Does it still have the will to build what it takes? Not simply talk about it, or simply regulate it. But build. Compete at the frontier. Take the risks, make the investments, and do the difficult things that leadership actually demands. Today, access to frontier AI still comes with painful tradeoffs. Costs that compound with every user. Latency that slows down products. Compute concentrated in a handful of providers. And too little control over where intelligence runs and where data lives. Why should building with the world’s best intelligence mean accepting these constraints? We don’t think it should. That’s why we’re building alpha.sh. Open models. Exceptional performance. Radically lower costs. Sovereign infrastructure. All accessible through one API. Europe has an extraordinary history of discovery. Of people who questioned what everyone else accepted, built what didn’t exist, and changed what the world believed possible. Is that spirit of discovery, invention, and ambition still alive today? Because in this new age, infrastructure is more than compute, networks, and data centers. It’s what gives a small team with an extraordinary idea a fighting chance against the largest companies in the world. It’s what turns the world’s best intelligence into new products, businesses, scientific discoveries, and capabilities that matter. It’s also what gives our businesses and institutions the ability to remain resilient, independent, and in control as intelligence becomes critical to everything they do. Europe needs this infrastructure. And it needs companies willing to put it to work. More founders building at the frontier. More businesses adopting the best models, testing their limits, and turning new capabilities into real advantages. More infrastructure built here, serving the world. Build something that wouldn’t have been possible yesterday. The models are here. The talent is here. The opportunity is here. What are we waiting for? alpha.sh/blog/introducing-al…
🤖 Made with AI
16
112
6
381
21,122
Der.dev 🔥🛠️ retweeted
The gap between an average engineer and the best engineers was never quite 10x, but it was meaningful. AI has made the gap at least 100x.
296
219
58
5,722
313,563
Der.dev 🔥🛠️ retweeted
spent last 10 years scaling distributed compute using three fundamental principles: shard, increase bandwidth, reduce latency. turns out intelligence has the same problem. introducing alpha. alpha.sh/blog/introducing-al…
Intelligence is becoming abundant. Access to it should be, too. Today, we’re introducing Alpha. We’re building the inference infrastructure for a world where open intelligence becomes the default. Starting with Europe. alpha.sh/blog/introducing-al…
28
69
6
291
11,227
Der.dev 🔥🛠️ retweeted
A huge amount of the Anti-AI code sentiment massively overestimates the quality of human code outside of a very small set of open source and high quality company codebases. Human Slop is everywhere and can trivially be improved on by any opus level model.
478
901
299
11,479
1,049,688
Der.dev 🔥🛠️ retweeted
"Did you know you can reduce tokens used by 6% by jumping through a bunch of hoops, making your life worse and also making your agents dumber? Why isn't everyone doing this!" - Half of current "startups"
144
69
20
3,035
89,273
Der.dev 🔥🛠️ retweeted
PSA: To use daybreak, sign in with your hardware key, not a random passkey. Otherwise, you will get cyber refusals due to improper authentication.
31
6
5
148
14,550
Der.dev 🔥🛠️ retweeted
Nice try cropping the image
Agreed. Roughly how I feel about desktop code app quality right now (admittedly slightly biased)
41
9
5
842
126,293
Der.dev 🔥🛠️ retweeted
hot take: you don't need a software factory, and i don't love the term. it's just marketing all you need is good software engineering and investing your tokens in the right places. the more deterministic the better (and token efficient)
all the companies selling you a software factory produce a shockingly tiny amount of software
432
569
178
8,000
467,904
Der.dev 🔥🛠️ retweeted
I can confidently say that humans were not very good at writing secure infrastructure. The endstate will be hypersecure infrastructure and software for all, but we're going to go through some turbulence before we reach that.
1,047
629
235
14,711
1,348,175
Der.dev 🔥🛠️ retweeted
We submitted a ~100-page full report on our assessment to Anthropic. For the public version, Anthropic redacted a majority of the evidence we found most informative. As a consequence, the public must rely on our judgment more than we consider ideal.
40
84
86
1,506
363,932