Bee0l@⚡🌋

trusted third parties are security holes
#Bitcoin

[Panta @thepanta82]

@theo There's a strong assumption here that tests can fully describe a system. But I've yet to make a user facing system where that's the case.

[Osaretin Victor Asemota @asemota]

My disillusionment with AI peaked this week as I attempted to perform complex analyses of real-life situations and model potential outcomes. I encountered several limitations rooted in:

1. A lack of full understanding of context while using natural language.

2. Limited availability of African data and relevant use cases.

The AI kept producing templated responses, which I systematically challenged. Eventually, it admitted that it didn't have the answers. While I felt relieved by this acknowledgment, I was also saddened by the realization.

Significant time and resources have been invested in platforms that are, at best, glorified calculators. I've come to understand what AI is suitable for and what it is not.

It's wise to avoid using AI to prepare for high-stakes interactions with others. However, it can be helpful for refining outputs and generating specific inputs. Ultimately, it is not true intelligence; it is merely data manipulation.

[Godwin Benson @busybenson]

A nation that is serious about developing its human capital through education must first be seen to be serious about the quality of teachers and teaching at all levels of education delivery.

This is a step in the wrong direction. Even the Holy Book says “teachers are worthy of stricter judgment”

Let’s not further lower the standard of the already embattled state of education in the country!

[Godwin Benson @busybenson]

37.5% and 25% cannot be acceptable cut off points in any exam globally.

Why are we lowering the standard @JAMBHQ?

[Godwin Benson @busybenson]

@tuteriacorp pays within this range for engineering talent.

If you��re great at problem solving, engineering and building with AI, send a DM. I’m actively recruiting Senior and Mid level engineers.

Backend & Full stack Engineers

Typescript-based.

[Bee0l@⚡🌋 @Beee_sama]

@trevorlasn @Dominus_Kelvin Stay away from server less database offering. The pricing is unpredictable. It isn't too difficult to setup postgres with backups to cloud flare/aws all running behind tailscale/wireguard. Claude/codex could even help with this.

[Andras Bacsai @heyandras]

https://t.co/NfSHagmEVG

[Jos Graas 🇱🇺 @graasdev]

@nextjs lol

[Anup Malani @anup_malani]

Why this matters for crime policy:

Most interventions target foot soldiers (arrests, deterrence, job training). But if it's a tournament, the policy lever is: reduce returns to winning.

Target gang leaders, not foot soldiers. Collapse the prize structure.

[ThePrimeagen @ThePrimeagen]

@DavidKPiano Non-coding architects have been one of the biggest banes in the programming industry of all time.

People claim they have met ones that are good, I have never met one that is good and I would be happy to work for

[banteg @banteg]

sorry but opus 4.6 is a completely retarded and braindead model. i only tolerate it because it writes better prose, but the levels of sloppiness and reward hacking are through the roof.

[Nathan Flurry 🔩 @NathanFlurry]

🧪 Experimental: Use OpenCode with Claude Code, Codex, and Amp

- Universal coding agent control
- HTTP API for sandboxed agents
- OpenCode TUI, web UI, SDK

Available in Sandbox Agent SDK 0.1.6

[Cody Caillet @codycaillet]

@MichaelAsiedu_ Without users and sales those are meaningless

[K.O.O @Dominus_Kelvin]

African startups should stop defaulting to AWS and Kubernetes.

Get a VPS on Hetzner, deploy with Coolify, and start experimenting until you start making money.

Oh, and when you do start making money, still stay there!

[Erik Schluntz @ErikSchluntz]

Next iteration of the Vibe Coder's Keyboard - now with Voice Mode!

[Cindy Sridharan @copyconstruct]

Unpopular opinion:

Unless you’re just prototyping, you should aim to understand as close to 100% of production code generated by LLMs.

Yes, all of it.

Effective mental models are still important for humans to sustainably maintain and evolve a codebase via prompting alone.

[S Max Brown @smaxbrown]

@Austen Indeed.

[James Long @jlongster]

@flybayer this is like waterfall development though. I just don’t buy into this :) it doesn’t work, as you implement something you want to continuously refine and change it

[James Long @jlongster]

I haven’t actually look at the ralph stuff

something about it just feels wrong to me. like a new level of slop by just getting whatever works

I still don’t understand why you don’t want to be deeply involved in the creative tech process

[Bee0l@⚡🌋 @Beee_sama]

@skeptrune This hasn't been my experience. I iterate faster in cursor than in cc. It seems easier to quickly course correct an agent when it appears to be going off in cursor bcos there is better visibility. And most importantly, you don't have to worry about getting banned unlike in cc