This is what using MANY (@many_markets) looks like: rates imported from the ECB, a EUR/USD butterfly built from a template, the agent asked to stress it, and the same structure replayed through real market history
Silas Asamoah retweeted
I am super excited 🎉 I'll be speaking on the Future of Software Development panel at Black Tech Fest next Thursday, 29 October, 4pm on Stage 2 at Tobacco Dock, London💜
Use my code GBADAGO10 for 10% off: joinbtf.com/attend/tickets
I've spent over seven years in tech, as a software engineer (full stack and backend), a developer advocate, a community builder and Vice President of a global open source foundation.
From building AI apps and delivering technical AI workshops to judging hackathons (Microsoft, UN, Cerebral Valley etc) on global tech stages, I've seen how the way we build software has changed so much in just the last few years. AI can now generate, review and test code, which is super exciting, but it also raises real questions about quality, security and what our role as engineers looks like next.
That's what we'll be talking about on the panel, including what's changed for engineers, the risks that come with it, whether developers will still be in demand and so much more.
If you're thinking about how AI is changing the way we build software, let's talk. DM me if you'll be there and come say hi after the panel!
NB: This isn't a paid post, I bought my VIP ticket months ago. Cmon! You know I invest in the things I'm passionate about! 😅
#BTF2026 #SoftwareEngineering #AI
MoMo scam calls and texts are incredibly annoying and in this era i figured why not use RL to catch the bad actors
I built sika-agent, an open RL environment for training AI agents that refuse these scams in English, Twi and especially Twi–English, because that mix is how I hear most people actually talk.
On results: Qwen3.5-9B completes 86% in English and 51% in Twi. The frontier models I ran stay at 90% or above in every language (about 240 tasks per model). That gap is what training small, local-language agents here should close.
The environment is available and open on the @PrimeIntellect env hub
Env Hub: app.primeintellect.ai/dashbo…
Code: github.com/Silas-Asamoah/afr…
Feedback/contibutions are welcome! I'll keep expanding this to include use cases I find interesting or useful
Thank you, @willcb! The credits that ran these evals came from you during the environment bounty program. I never submitted back then so this is my submission, finally 😄
Silas Asamoah retweeted
Okomfo Anokye is next, cleaning up animations for anokye character. The character be 🔥🔥
#TribalKombat #GameDev #IndieGameDev #Devlog
Silas Asamoah retweeted
This was definitely the moment when I realized I was playing the greatest video game of all time
Silas Asamoah retweeted
I started out as a @wtmaccra community member. Now I’m a @WomenTechmakers Ambassador, one of 512 selected from 1,900+ applications across 130 countries. Still sinking in.
If you’re a woman trying to break into tech, my DMs are open.
#WomenTechmakers #WTMAmbassador #WomenInTech
Silas Asamoah retweeted
Ten years of Sad Affleck memes and it turns out he'd just been trying to install CUDA drivers the whole time.
Silas Asamoah retweeted
Jeff Dean leaving Google and Gwern breaking pseudonymity are 2/7 seals of the apocalypse
Silas Asamoah retweeted
This guy is single handedly representing Ghana in a sport we don’t even hear about and I’ve seen nary a news outlet give him 2 minutes of coverage. He’s self-sponsored & gives his all but we don’t even appreciate him. I’m gonna start following his every activity. Congrats Trey.
I guess you can say there’s now RLM at home
Replying to @PrimeIntellect
Prime Agent is a general-purpose coding harness
On ARC-AGI-3, it scores 95.5%, surpassing the human-expert baseline, but the gain is not benchmark-specific.
We see major improvements across models when compared to their proprietary harnesses:
Silas Asamoah retweeted
Introducing Prime Agent:
A self-improving RLM harness for coding and long-running autonomous tasks.
Designed to be both token-efficient and expressive through programmatic tool calling, context as a variable, multi-agent messaging, and a self-modifiable harness state.
























