Andrei Onel
@AndreiOnel
Followers
277
Following
281
Media
19
Statuses
274
founder @askmanu_ helping companies document their code base
Bucharest / Dublin
Joined December 2011
I'll write more about the things learned there but if you can, I recommend you join the next batch starting Oct 6. It's the best investment you can do for your team.
maven.com
Learn proven approaches for quickly improving AI applications. Build AI that works better than the competition, regardless of the use-case.
0
0
0
The course has already improved @askmanu_ dramatically and I'm launching a new product super soon that has been made 10x better by evals.
1
0
1
I was happy to catch a spot in @HamelHusain and @sh_reya most recent Maven course on AI evals. The course just finished last week and wow. My eyes are open. I just left the cave. The shadows on the wall make sense now.
1
0
0
AI evals are the moat. If you're building in the AI space, evals are starting to become the most important thing and in most cases the only thing that distinguishes products from demos.
1
0
0
RT @willccbb: @sammcallister @aidan_mclau would this also apply to employees at Google? Meta? xAI? Mistral? Cohere? Cursor? Windsurf? Lovab….
0
2
0
This whole thing leaves a really bad taste in my mouth. So no one can use claude to work on AI products?.
@aidan_mclau This take is sad to see but you might not have full context. We cut OpenAl's access for violating our APl terms and for the heavy usage of Claude Code among OAI tech staff. We're going to continue providing API access for safety evals and benchmarking. That's important to us.
0
0
0
RT @tech_optimist: It's HERE! First-class rewrite (not port) of @DSPyOSS in Rust 🦀. I know I'll be watching its development and playing a….
github.com
A DSPy rewrite to(not port) Rust. Contribute to krypticmouse/DSRs development by creating an account on GitHub.
0
14
0
RT @deedydas: China (yet again) just dropped the best open weight coding model. Qwen3-Coder beats Kimi K2, which came out 2 weeks ago, on….
0
81
0
I've just discovered @dylan522p and I think I watched every interview he gave in the past 6 months. It's like drinking from a fire hose. Amazing energy and knowledge. We need him on every podcast.
0
0
1
RT @HamelHusain: In our rage against "RAG is Dead", @bclavie and I assembled annotated talk notes (Simon Williston style). We take you th….
hamel.dev
Why the future of RAG lies in better retrieval, not bigger context windows.
0
21
0
Super happy to announce that @peermetrics has been acquired by @WebRTCVentures. There is no team better equipped to continue the work of creating the most complete WebRTC monitoring tool out there. Check out the project on github.
github.com
A self-hosted, open source WebRTC analytics platform for monitoring and optimizing real-time communication apps. Now under the stewardship of WebRTC.ventures. - Peermetrics
0
0
0
Super surprising to see BM25 outperform OpenAI text-embedding-3-small.
fresh-stack.github.io
FreshStack: Building Realistic Benchmarks for Evaluating Retrieval on Technical Documents.
0
0
2
Great presentation from @beirmug yesterday about RAG evals. FreshStack looks like the exactly what the next evolution of IR evals should be. Great stuff.
Existing IR/RAG benchmarks are unrealistic: they’re often derived from easily retrievable topics, rather than grounded in solving real user problems. 🧵Introducing 𝐅𝐫𝐞𝐬𝐡𝐒𝐭𝐚𝐜𝐤, a challenging RAG benchmark on niche, recent topics. Work done during intern @databricks 🧱
2
0
13
RT @theinformation: Exclusive: OpenAI is mimicking Palantir in customizing AI models for customers spending $10 million or more. Read more….
theinformation.com
As OpenAI pitches its artificial intelligence to big companies and government agencies, it’s taking a page from the playbook of many software firms by offeringin-house researchers and software...
0
36
0
RT @andreasklinger: EU-Inc is close to the finishing line (of the first stage) 🇪🇺. All the politicians in Brussel want "something like it".….
0
158
0
RT @lateinteraction: DSPy's biggest strength is also the reason it can admittedly be hard to wrap your head around it. It's basically say:….
0
150
0
RT @RubenHssd: BREAKING: Apple just proved AI "reasoning" models like Claude, DeepSeek-R1, and o3-mini don't actually reason at all. They….
0
10K
0
RT @vitrupo: OpenAI's Greg Brockman says the AGI future looks less like a monolith - and more like a menagerie of specialized agents. Model….
0
156
0
RT @ziv_ravid: You know all those arguments that LLMs think like humans? Turns out it's not true. 🧠 In our paper "From Tokens to Thoughts….
0
322
0
Talking to Dario feels like talking to a super smart geeky person. Really great listening to him.
0
0
0