
Mbongeni Ndlovu
@Mbounge_
Followers
4K
Following
71K
Media
2K
Statuses
36K
Computational Sports Scientists, MSc, Strength&Conditioning, Olympic Weightlifting Coach
Canada, Nova Scotia
Joined August 2023
RT @alignment_lab: Introducing SENTER.We are announcing the availability of SENTER, a powerful workstation we built to perform research and….
0
17
0
RT @dromanocpm: The most beautiful part of having integrity is there literally isn’t a thing or person you should be unwilling to walk away….
0
8
0
If no reasoning model leads . Why are the labs slowing down on releasing non-reasoning models.
Agent Leaderboard v2 is here!. > GPT-4.1 leads.> Gemini-2.5-flash excels at tool selection.> Kimi K2 is the top open-source model.> Grok 4 falls short.> Reasoning models lag behind.> No single model dominates all domains . More below:
0
0
0
Just intermittent fast instead . 16 hrs no food . 8 hr window to eat. Only eat 2-3 hours before bed and go through the morning and afternoon without food. I find I'm more focused during the day like this. The moment food enters my body . Focus goes down.
Dane and I agreed to do a fast after launching Grok 4. 72 hours of no food later, I can confidently say: 0/10 would not recommend, this was needless suffering. Surprisingly I got a lot of work done, but ketosis clarity is pure fiction. All health influencers are capping
1
0
1
Love this Ryan ♥️. About 2 weeks ago, I was helping a friend with an interview he was preparing for to be an Assistant Strength Coach at a university. He had never interacted at all with anything AI (ChatGPT etc. ). We sat down and started talking through what he wanted to show.
had a 45 min convo with my uber driver this morning on the way to the airport. he asked what i do and we went down the ai rabbit hole. he was a truck driver that converted to uber because he couldn't break even on the lease payments for his rig. we talked about what's going to.
1
0
3
Reading between the lines of this post, it becomes clear - each country or culture must train its own models. The homogeneity of Western AI influence runs deep in the global psyche.
Kimi has a distinct writing style that is free of most of the patterns we now associate with AI generated text. Both Kimi and DeepSeek's prose is apparently even more impressive in Chinese. Both of these models have a unique 'voice', quite different from Western AI.
0
0
2
Limitation of LLMs.
Something about this kind of prompt is simply unfathomable to LLMs. They just can't perform better than chance, and I'm not sure why. Most people will dismiss this as just being "hard math stuff", but it is not, I swear. It is just alien to you because it is *niche*, thus, it
0
0
0
Too early to say . Labs have been benchmaxing too much. Need the model to be tested by lots of people in the public to get a true understanding of how the model operates in the real world.
XAI GROK 4 BENCHMARKS:. > openai o3 is cooked .> gemini 2.5 pro is cooked .> claude opus 4 is cooked . ITS OVER, GROK 4 WON
0
1
2
I hope "context engineering" catches on.
+1 for "context engineering" over "prompt engineering". People associate prompts with short task descriptions you'd give an LLM in your day-to-day use. When in every industrial-strength LLM app, context engineering is the delicate art and science of filling the context window.
0
0
2