b0bleet

@nexxeln doing useless things

why is gpt-5.5 being so dumb rn

@Kyler_Lorin yes it is very bad especially for coding tasks,

Did Openai do something to GPT 5.5? It's acting different when coding.

@thsottiaux it performs very bad for a few days.

Codex team is aware of reports of GPT-5.5 performing worse for some users and investigating. We don't have anything conclusive yet and systems are healthy but we will share updates as we go.

@jorandirkgreef I believe that and follow/like TigerBeetle approaches, but I understand Jarred pains to maintain bun with zig.

@bobl33t We work hard to keep TigerBeetle simple.

We’ve had on the order of 3 memory bugs in 6 years of TigerBeetle. None RCEs. On the other hand, our own simulators have proactively found hundreds of (devastatingly catastrophic) distributed systems correctness bugs per year. Given how hard TigerBeetle’s domain is, in terms of mission critical financial transaction processing, I’ve never for one minute believed that writing TB in a memory safe language such as say TypeScript would somehow magically (!) make any material impact compared to the 100x correctness multiplier of TigerStyle. That’s because—rather than fall for the fallacy of composition, i.e. to see distributed correctness as a language problem—TigerStyle instead takes ultimate responsibility for the “end to end” correctness of the distributed system as a whole. Per systems engineering, correctness is always a systems design problem. For example, how to build a reliable whole, (especially) out of unreliable parts, such as broken firmware, bitrot, programmer error etc. In other words, application of the end to end principle. But when you TigerStyle the design in this way, the world of systems engineering also completely opens up to you and changes how you evaluate systems languages (now things like “power to grammar ratio”, or explicitness, checked arithmetic and precision become more critical and valuable to you). Of course, it is harder to care about correctness, to take responsibility for correctness end to end. Yes, you’re forced to begin to worry about the more serious concerns, starting with the basics of static allocation, explicit limits, assertions, deterministic simulation testing and moving to more advanced topics like protocol-awareness and storage fault-tolerance. But then again, TigerStyle is such a force multiplier, that you achieve mission critical quality, and in less time and with greater velocity. If you’re tired of production issues, and if you want to “engineer your engineering”, I would encourage you to lift up your thinking to the level of systems design and end to end correctness. Start thinking about your methodology and begin embracing TigerStyle. tigerstyle.dev

@sama @rezoundous gpt 5.5 nerfed. i feel it for a few days.

@rezoundous curious, could you try to explain it?

Can't explain it, but I trust GPT-5.5 more than Opus 4.7 right now.

@Shivam25mishra mac os

Linux users be honest. Ubuntu or Fedora?

@abhitwt we migrated from postman (team plan) to bruno

@flpsnd i switched into codex from claude because claude code is very bad and don't follow my rules.

i don't want to be a hater, but claude code feels really weird lately, like i'm amazed by how much it tends to fuck up shit

@bcherny @tengyanAI I think quality dropped in practice. Example: it now generates fmt.Println("...\n") even though Println already adds a newline. Also seeing functions returning error without err != nil checks. These regressions weren't there before.

This is false. We defaulted to medium as a result of user feedback about Claude using too many tokens. When we made the change, we (1) included it in the changelog and (2) showed a dialog when you opened Claude Code so you could choose to opt out. Literally nothing sneaky about it — this was us addressing user feedback in an obvious and explicit way.

basically: anthropic sneakily turned down how hard claude thinks before editing code, changed the default from "high" to "medium" effort, and hid the reasoning from session logs. all without telling users. an amd director had 7k sessions of telemetry to prove the degradation was real and measurable (not just vibes). anthropic admitted to the changes. there's a workaround (use "/effort max"). the uncomfortable part is most users had no data to notice it happened at all.

@thekitze same nowadays it is very dumb.

cc is over, only for copers and fanbois

@trq212 does it use webmcp?

POV: you're cooking

@pcshipp m4 pro max

Which MacBook should I choose for coding? MacBook Air (M1) MacBook Air (M2) MacBook Pro (M3) MacBook Pro (M4 Pro)

@glcst small query language

What does SQL mean ? Wrong answers only

@vikhyatk i also switched to kitty as well from ghostty but battery usage wasn't too bad.

switched from ghostty to kitty battery life increased by at least 20%

@tm23twt thanks great blog. i built gpt-2 with tinygrad: github.com/b0bleet/teenypt next plan is to write blog about that

blog on - Learnings after coding GPT & Llama from scratch :) this contains almost everything that i learned through the process, some very fundamentals yet important & subtle points that you should keep in mind. just a 4 min chill read, link in replies✌️

@CodeByNZ depends on language

How do you name variables?

i have implemented gpt-2 with tinygrad. it was challenging, but I learned a lot along the way. github.com/b0bleet/teenypt

@nizzyabi linear is the best

do any newer startups use jira?

@fidexcode i have both macbook and thinkpad x1 carbon gen13 (with good battery life) but still prefer using macbook pro for development.

Let's end this debate Thinkpad or Macbook