AI developer | Cybersecurity researcher | Technical writer | Exploring the systems shaping AI, security, and digital policy.
Nigeria
- Tweets 240
- Following 613
- Followers 107
- Likes 174
Muhammad Abdullahi dattijo retweeted
What, exactly, is the edge that agentic RAG has over regular RAG? @DattijoMuhammad sets out to find out through extensive testing — which produced counterintuitive results. towardsdatascience.com/does-…
1
17
3,147
Muhammad Abdullahi dattijo retweeted
Best 15 accounts to follow in AI:
@karpathy = LLMs king
@steipete = built openclaw
@gregisenberg = startup ideas king
@rileybrown= vibecode king
@jackfriks = solo apps king
@levelsio= startups king
@marclou = startups king
@EXM7777 = AI ops + systems king
@eptwts = AI money twitter king
@ai_explorer25= AI queen
@godofprompt =prompt king
@vasuman = AI agents king
@AmirMushich = AI ads king
@0xROAS = AI UGCs king
@egeberkina = AI images king
Follow them all and learn.
34
129
765
92,996
Muhammad Abdullahi dattijo retweeted
How well do AI image judges perform, and what are the areas where they still struggle?
@DattijoMuhammad explores a complex set of challenges in his new deep dive on vision AI models. towardsdatascience.com/i-bui…
1
5
2,188
GPT-4o fell for a different one. On the tilt question, where neither image was tilted more and the right answer was SAME, it picked an image 72.5% of the time, often sounding confident. Claude did that 17.5% of the time, Gemini 28.3%.
1
1
19
Caveats: small test, artificial differences, Gemini 3.1 Flash Lite instead of the main Gemini model (quota), and not comparable to VisionQ scores. Full article: towardsdatascience.com/i-bui…
1
1
16
I tested four ways to give an AI assistant a memory. The obvious one, pasting the whole history into every prompt, was the worst value: 20 to 31 times the cost of retrieval, and it did not win on accuracy either. Thread:
1
38
Full history also struggled with corrected facts: mini completed 46% of tasks on corrected facts vs 66% on unchanged ones. The running summary lost facts too: mini kept the newest value for 20 of 30, Haiku for 28.
1
4
Retrieval completed 88.9% of tasks on mini and 85.6% on Haiku, and fetched every needed piece in 93% of runs. Caveat: 30 tasks, one made-up person, so read it as what to test, not a verdict. Full article: towardsdatascience.com/i-tes…
6
I made an apartment search agent 25x cheaper and it still found all 101 true matches with zero false positives. Cost for 500 listings x 5 requirements went from $0.20 to $0.008. Sending every pair to the strong model would have cost $1.88. Here is what actually moved the bill.
16
14
579
Plot twist: giving the strong model LESS text helped. On one hard listing, the full listing prompt got it right 0 of 10 times. Title and body only got it right 10 of 10.
1
5
328
Final setup: a cheaper model with the strong model as backup. 129 calls, $0.008, all 101 matches, zero false positives. Caveats: one run per setup, and I picked the escalation threshold after seeing results. Full write-up: towardsdatascience.com/can-a…
224
3
167
885