Andrew Schwabe (@aschwabe) 's Twitter Profile
Andrew Schwabe

@aschwabe

Hacker entrepreneur; Artificial Intelligence researcher

ID: 25590349

linkhttps://formotiv.com calendar_today20-03-2009 22:34:01

2,2K Tweet

557 Takipçi

179 Takip Edilen

Andrew Schwabe (@aschwabe) 's Twitter Profile Photo

Truthy vs. Creative AI: Classic AI is reproducible & explainable—perfect for precision. GenAI, driven by probabilities & “temperature,” excels in creativity but lacks consistency. Know your use case, test rigorously, & monitor results to leverage AI effectively! #AI #TechTrends

Truthy vs. Creative AI: Classic AI is reproducible & explainable—perfect for precision. GenAI, driven by probabilities & “temperature,” excels in creativity but lacks consistency. Know your use case, test rigorously, & monitor results to leverage AI effectively! #AI #TechTrends
Andrew Schwabe (@aschwabe) 's Twitter Profile Photo

AI privacy is shaky—data leaks are real. Run massive models locally with quantization! It shrinks model size & boosts speed by grouping weights like LEGO bins. K-Quant & Context Quantization make LLMs edge-friendly. Thoughts on AI optimization? ☕ #AI #Quantization

AI privacy is shaky—data leaks are real. Run massive models locally with quantization! It shrinks model size & boosts speed by grouping weights like LEGO bins. K-Quant & Context Quantization make LLMs edge-friendly. Thoughts on AI optimization? ☕ #AI #Quantization
Andrew Schwabe (@aschwabe) 's Twitter Profile Photo

Vietnam's new data protection law (Decree 13/2023/ND-CP) is live! 🛡️ Covers all data processors, mandates clear consent, and grants strong user rights. Businesses, ensure compliance to avoid penalties! 🔒 Details: vietanlaw.com/decree-13-2023… #DataPrivacy #VietnamLaw

Andrew Schwabe (@aschwabe) 's Twitter Profile Photo

Customer service is dead. Unsolicited emails are now standard, with drip campaigns guilting you for not responding. Unsubscribe links make you re-enter your email—like they can't auto-fill it after 20+ years of tech. AI's here, but we can't even handle email lists responsibly.

Customer service is dead. Unsolicited emails are now standard, with drip campaigns guilting you for not responding. Unsubscribe links make you re-enter your email—like they can't auto-fill it after 20+ years of tech. AI's here, but we can't even handle email lists responsibly.
Andrew Schwabe (@aschwabe) 's Twitter Profile Photo

edition.cnn.com/2025/08/26/tec… We are NOT at a level where any AI products are intelligent. Just because they answer some things correctly and automate some tasks does not demontrate empathy, understanding, emotional intelligence, or any reasonable value of life like a human can.

Andrew Schwabe (@aschwabe) 's Twitter Profile Photo

This is an excellent article that highlights managerial patterns of how AI's benefits are devolving discipline. Anybody who is building product, take a good look. None of these are intentional, but a rising 'executive pressure' is the open door. buff.ly/0diYpGD

Andrew Schwabe (@aschwabe) 's Twitter Profile Photo

Doubling down on classic AI in my new HackerNoon article: Why science beats spectacle for business success. Reliable, explainable methods over hype—focus on real results! 📈 hackernoon.com/doubling-down-… #AI #Tech #Business

Doubling down on classic AI in my new HackerNoon article: Why science beats spectacle for business success. Reliable, explainable methods over hype—focus on real results! 📈 hackernoon.com/doubling-down-… #AI #Tech #Business
Andrew Schwabe (@aschwabe) 's Twitter Profile Photo

LLMs are to Truth what "voter fraud" is to elections. Do you agree? disagree? Lets get a respectful argument going! I am talking about trust as it pertains to intelligence.

LLMs are to Truth what "voter fraud" is to elections.  Do you agree? disagree? Lets get a respectful argument going!  I am talking about trust as it pertains to intelligence.
Andrew Schwabe (@aschwabe) 's Twitter Profile Photo

🚨 Wake-Up Call: 5 GenAI Misconceptions Debunked 🚨 Think AI is the "universal solver" or path to AGI? Think again. This HackerNoon piece cuts through the hype: ❌ LLMs aren't magic. Use wisely! hackernoon.com/calling-all-br…

🚨 Wake-Up Call: 5 GenAI Misconceptions Debunked 🚨
Think AI is the "universal solver" or path to AGI? Think again. This HackerNoon piece cuts through the hype:
❌ LLMs aren't magic. Use wisely!
hackernoon.com/calling-all-br…
Andrew Schwabe (@aschwabe) 's Twitter Profile Photo

Overwhelmed by AI? From FOMO to constant distractions, it's real—but manageable! Set boundaries, prioritize tools, take detox breaks, and remember: your humanity trumps tech. Reclaim control!: hackernoon.com/how-to-navigat… #AIOOverwhelm #TechWellness #AIAnxiety

Overwhelmed by AI? From FOMO to constant distractions, it's real—but manageable! Set boundaries, prioritize tools, take detox breaks, and remember: your humanity trumps tech. Reclaim control!: hackernoon.com/how-to-navigat… #AIOOverwhelm #TechWellness #AIAnxiety
Andrew Schwabe (@aschwabe) 's Twitter Profile Photo

🔥 NVIDIA drops Nemotron 3: the most efficient open model family for agentic AI 3 sizes (Nano/Super/Ultra) with breakthrough hybrid MoE architecture Nano specs: 4x throughput vs Nemotron 2 60% lower reasoning costs 1M token context Available NOW on HuggingFace

🔥 NVIDIA drops Nemotron 3: the most efficient open model family for agentic AI
3 sizes (Nano/Super/Ultra) with breakthrough hybrid MoE architecture
Nano specs:

4x throughput vs Nemotron 2
60% lower reasoning costs
1M token context
Available NOW on HuggingFace
Andrew Schwabe (@aschwabe) 's Twitter Profile Photo

MacOS Tahoe experiment: FAILED. Unclear system update pushed me from Sequoia → Tahoe. Result: 5s boot → 2min, app compatibility nightmare, manual restore hell (Apple blocks newer→older migration). Even AI hallucinated about the migration process. Staying on Sequoia for now.

Andrew Schwabe (@aschwabe) 's Twitter Profile Photo

GenAI is impressive for images/video/text, but it's not intelligence—it's probabilistic math. Token prediction ≠ understanding. Hallucinations are inevitable. We must move focus from scaling LLMs to cognitive AI that enables true synthetic understanding and real learning.

GenAI is impressive for images/video/text, but it's not intelligence—it's probabilistic math. Token prediction ≠ understanding. Hallucinations are inevitable. We must move focus from scaling LLMs to cognitive AI that enables true synthetic understanding and real learning.
HackerNoon | Learn Any Technology (@hackernoon) 's Twitter Profile Photo

Alignment-Driven Development is the governance layer that AI-accelerated teams are missing — a documented, auditable process that keeps every product decision - hackernoon.com/alignment-driv… #artificialintelligence #aigovernance

Andrew Schwabe (@aschwabe) 's Twitter Profile Photo

The Facebook feed is ironically no longer full of faces I know. Just ads and algorithm bait — and somehow the settings to fix that "disappeared." 🙃 Hot tip: want to mess with the algorithm? Follow the weirdest rabbit trails you can find on YouTube, Facebook. Watch chaos unfold.

The Facebook feed is ironically no longer full of faces I know. Just ads and algorithm bait — and somehow the settings to fix that "disappeared." 🙃
Hot tip: want to mess with the algorithm? Follow the weirdest rabbit trails you can find on YouTube, Facebook. Watch chaos unfold.
Andrew Schwabe (@aschwabe) 's Twitter Profile Photo

Honored to receive an Emerging Scholar Award at the 19th International Conference on e-Learning & Innovative Pedagogies 🎓 Focus this year: Human-Centered AI — which is exactly what we're building at Saigon A.I. ubi-learn.com/2026-conferenc… #HumanCenteredAI #eLearning

Honored to receive an Emerging Scholar Award at the 19th International Conference on e-Learning & Innovative Pedagogies 🎓

Focus this year: Human-Centered AI — which is exactly what we're building at Saigon A.I.

ubi-learn.com/2026-conferenc…

#HumanCenteredAI #eLearning