@gadbenram

โœก๏ธ โ€ข Founder of @tensoropsai โ€ข Your partner for AI ๐Ÿค–

 Lisbon, Portugal
tensorops.ai
Joined September 2016
  • Tweets393
  • Following261
  • Followers266
107Media
ื”ืจืฆืืช ื”ืคืชื™ื—ื” ืฉื”ืขื‘ืจืชื™ ื‘ื›ื ืก AI ืœื›ืœ ื’ื•ืคื™ ื”ืชืงืฉื•ืจืช ื”ื—ืฉื•ื‘ื™ื ื•ื”ื’ื“ื•ืœื™ื ื‘ืกืคืจื“
3
1
22
76,046
ื”ื›ื ืก ื”ืžืœื: youtube.com/watch?v=l-FHlO1iโ€ฆ

II Foro Iberoamericano de IA Aplicada al Periodismo - 30 de septiembre

Innsomnia Group celebra el prรณximo 30 de septiembre y 1 de octubre ...

youtube.com
1
129
Mr @POTUS, I'm glad to share with you that we registered under the tensorops.si domain!
๐Ÿค– Made with AI
1
2
81
Gad Benram retweeted
Gad Benram
@gadbenramSep 19
Replying to @Reuters
LEAKED: Screen recording of the massive Gemini hack
๐Ÿค– Made with AI
1
20
Shana Tova everyone! And thanks @OpenAI for the new partnership this year!
1
111
ื›ืฉืฉืžืขืชืชื™ ืฉืžื’ื–ื™ืŸ ื˜ื™ื™ื ื‘ื—ืจ ืืช ืคืจื™ืก ื”ื™ืœื˜ื•ืŸ ืœืื—ื“ ืž-100 ืžืฉืคื™ืขื ื™ ื” AI ื”ื—ืฉื•ื‘ื™ื ื‘ืขื•ืœื ืื‘ืœ ื’ืณื ืกืŸ ื”ื•ืื ื’ ืœื ื ื‘ื—ืจ
Haimon
@Haimon100Aug 27
ื‘ืขืงื‘ื•ืช ื”ื—ืชื•ื ื” ืฉืœ ืขื•ืžืจ ืื“ื ื–ื” ืžืชื‘ืงืฉ ๐Ÿ‘‡๐Ÿผ
1
3
136
Gad Benram retweeted
Gad Benram
@gadbenramAug 25
Replying to @WSJ
โ€œClaude, draft my IPO prospectus. Make no mistakes.โ€
1
1
104
Gad Benram retweeted
Zach Moskow
@zachmoskowAug 15
Thanks @OpenAI for the limited edition merch
636
1,755
588
51,127
5,218,791
Zionism 1 was crazy.
๐™๐๐ค โ€”
@zadokismAug 11
Zionism 1 was crazyyy
38
102
5
2,233
311,808
Gad Benram retweeted
Eylon Levy
@EylonALevyJul 18
Tel Aviv just before sunset โค๏ธ
116
200
24
2,444
155,035
GLM 5.2 just one shotted "Super Plumber" and the price to intelligence amazing! Finally there is an Open Source model at the level of Opus and the price is about 90% less in price. I got access to their token service from a company (currently under NDA) that is based on kernel-level optimization and other breakthroughs that should not be disclosed. But what I can tell you is this: I took their system for a ride. I plugged it into a code clue and made it think it was using Opus through an LLM Gateway that fakes the readings. Then I asked for a single shot to build me a "Super Mario", with no corrections, no feedback from me. 11:33 minutes run, 47.0K tokens, 1451 lines written and 34 lines deleted. The price if I had actually used Opus would have been $10.69, but I actually paid exactly $1.25. And all this is above the Open Source model. So, run urgently to check out GLM 5.2, and if you happen to spend over $15,000 a month on Claude for your organization (which according to my experiment came out 1500 a month...), maybe I'll be able to get you early access to the platform..
2
26
210,244
I'm proud to share a comprehensive and unique tutorial covering everything you need to know about loop engineering while .... [your code here] end
1
141
Prepping for your AI certs? ๐Ÿง  100% FREE practice exams for the AWS GenAI Developer (AIP-C01) & Claude Essentials are now live on TensorLearn! Get started here: tensorops.ai/academy #AWS #Claude #AI
1
1
181
ื”ืกืงื™ืœ ื”ื—ื“ืฉ ืฉื›ื•ืœื ืžื“ื‘ืจื™ื ืขืœื™ื•: Caveman ืžื’ื ื™ื‘, ื”ื•ื ื’ื•ืจื ืœ LLM ืœื“ื‘ืจ ื›ืžื• ื’ื‘ืจ ื˜ื™ืคื•ืกื™ ื‘ื–ืžืŸ ืฉื”ื•ื ืฆื•ืคื” ื‘ืžื•ื ื“ื™ืืœ ื•ื‘ื›ืš ืžื•ืจื™ื“ ืืช ื”ื˜ื•ืงื ื™ื ืฉื”ื•ื ืžืฉืชืžืฉ ื‘ื”ื. ืœืžื” ื–ื” ื‘ื›ืœ ื–ืืช ื‘ื“ื™ื—ื” ื•ืœืžื” ื–ื” ื‘ืขืฆื ื›ืŸ ื ื•ืฉื ื—ืฉื•ื‘? ืื– Caveman ื–ื” ืกืงื™ืœ ืฉื™ื’ืจื•ื ืœ LLM ืฉืœื›ื ืœื“ื‘ืจ ื›ืžื• ืื™ืฉ ืžืขืจื•ืช: ืดืื ื™ ื‘ืื’, ืžืชืงืŸ, ื˜ืกื˜ ืขื‘ืจ.ืด ื”ืคืจื•ื™ื™ืงื˜ ืคื•ืชื— ื›ื‘ื“ื™ื—ื” ื•ื‘ืคื•ืขืœ ื–ื” ืœื ืžืžืฉ ืขื•ื–ืจ ืขื ื”ื‘ืขื™ื” ื”ืืžื™ืชื™ืช ืฉืœ ืกื•ื›ื ื™ ืงื•ื“ ืžืื—ืจ ื•ืจื•ื‘ ื”ื˜ื•ืงื ื™ื ื”ื ืœื ื‘ืชืฉื•ื‘ื” ื”ืกื•ืคื™ืช ืœื™ื•ื–ืจ ืืœื ื‘ืขื•ื“ืคื™ ื”ืงื•ื“ ื•ื” reflection ืฉื ื•ืฆืจื™ื ื‘ืื™ื˜ืจืืฆื™ื•ืช ื”ืคื ื™ืžื™ื•ืช ืฉืœ ื” reasoning. ืื‘ืœ ื”ื•ื ื›ืŸ ืžืขืœื” ื ืงื•ื“ื” ื—ืฉื•ื‘ื” ืœื’ื‘ื™ ืžื•ื“ืœื™ ืฉืคื” ืฉื™ืฉ ืœื”ื ืžื•ื˜ื™ื‘ืฆื™ื” ืžื•ื‘ื ื™ืช ืœื”ื•ืฆื™ื ืคืœื˜ ื‘ื’ืœืœ ืื•ืคืŸ ื”ืื™ืžื•ืŸ ืฉืœื”ื ืฉื ื•ืชืŸ ืœื”ื ืชื’ืžื•ืœ ื—ื™ื•ื‘ื™ ืขืœ ืœื”ื•ืฆื™ื ืคืœื˜ ื›ืœืฉื”ื•, ืชื•ื›ืœื• ืœืงืจื•ื ืขื•ื“ ื‘ืžืืžืจ ืฉืœ Adam Tauman Kalai, Ofir Nachum, ืื‘ืœ ื”ื ื‘ื’ื“ื•ืœ ื”ืฆื™ืขื• ืฉื ืจืขื™ื•ืŸ ืฉื“ื•ืžื” ืœืžื‘ื—ืŸ ืกื•ื‘ื™ื™ื˜ื™ - ื ื™ื—ื•ืฉ ืฉืœ ืชืฉื•ื‘ื” ืœื ื ื›ื•ื ื” ืžื•ืจื™ื“ ื ืงื•ื“ื•ืช, ืื– ืื™ืŸ ืžื•ื˜ื™ื‘ืฆื™ื” ืœื ื—ืฉ ืกืชื ืื ืœื ื™ื•ื“ืขื™ื ืืช ื”ืชืฉื•ื‘ื”. ื‘ื›ืœ ืžืงืจื”, ืกืงื™ืœ ืžืฉืขืฉืข ื•ืงื•ื ืกืคื˜ ืžืขื ื™ื™ืŸ!
1
71
A few weeks ago, I wrote a post arguing that successful startups don't typically build "agents" as their primary product. Recently, I came across an interesting analysis that highlights how common perceptions about where the money is in AI can sometimes miss the mark. @chris__lu Wrote an overview that stated that 70% of the startups in the YC Spring 2026 batch are building agents. However, a closer look at the data methodology reveals a slightly different picture. Upon a second read, it appears the review simply checked for the occurrence of the phrase "AI Agent," assuming that any startup mentioning the word is actively building one. A more thorough examination, using Gemini to classify what these startups actually declare they are doing, found that out of 196 companies checked in the batch, only 38 to 61 are truly "building agents" (depending on the classifier's prompt definition). The rest are either building infrastructure for agents (like telephony or browsers for agents) or doing something else entirely.
1
2
90
Original post: xitter.cf/i/status/2066199289546โ€ฆ
Chris Lu
@chris__luJun 14

I analyzed all 196 YC Spring 2026 companies: here's what it tells us about the future.

Article ยท @chris__lu
2
79
ื‘ืฉื ืช 2024 ื“ื™ื‘ืจืชื™ ืžื•ืœ ื‘ื•ืจื“ ื•ื”ื ื”ืœื” ืฉืœ ื—ื‘ืจืช ื ืืกื“ืง ืฆื™ื‘ื•ืจื™ืช ืขืœ ื–ื” ืฉืชืคื™ืกืช ื”ืชื•ื›ื ื” ื”ื•ืœื›ืช ืœื”ื™ื•ืช ืฉื•ื ื” ืžื”ื•ืชื™ืช ื‘ืขืชื™ื“. ืงื•ื“, ื”ื•ืœืš ืœื”ื™ื•ืช ืžื™ื•ืฆืจ ืžืชื•ืš ืื™ื ื˜ืจืืงืฆื™ื” ืขื ื”ื™ื•ื–ืจ, ื‘ื–ืžืŸ ืืžืช. ื”ื™ื” ืžื™ ืฉื”ืกื›ื™ื ืื™ืชื™, ืœืื•ืจืš ื”ืฉื ื™ื ื›ืฉื”ืžืฉื›ืชื™ ืœื˜ืขื•ืŸ ืืช ื–ื” ื”ื™ื• ืžื™ ืฉื”ืชื™ื™ื—ืกื• ืœื–ื” ื‘ื ืคื ื•ืฃ ืฉืœ ืคื•ืคื•ืœื™ื–ื. ืื ื™ ืฉืžื— ืœืจืื•ืช ืืช ื”ืžืฆื™ืื•ืช ื”ื–ืืช ื ืจืงืžืช ืœืžื•ืœ ื”ืขื™ื ื™ื™ื ื•ื”ื ื” ืขื•ื“ ื“ื•ื’ืžื ืฉืืกื•ืจ ืœื›ื ืœื”ื—ืžื™ืฅ, ื‘ืžื™ื•ื—ื“ ืื ืืชื ืขื•ืกืงื™ื ื‘ Security ื•-AI. ืžืจ ื‘ื—ื•ืจ ืžื“ื’ื™ื ืื™ืš ื”ื•ื ื‘ื ื” ืืช VibeOS, ืžื” ืฉื’ื•ืจื ืœื›ื ืœืฆื—ื•ืง ื‘ื“ืงื•ืช ื”ืจืืฉื•ื ื•ืช ื”ื•ืคืš ืœืœืกืช ืฉืžื•ื˜ื”. ืžืขืจื›ืช ื”ื”ืคืขืœื” ื”ื–ืืช ื‘ื ื•ื™ื” ื›ืš ืฉื›ืœ ืืคืœื™ืงืฆื™ื” ื‘ื” ื ื•ืฆืจืช ื‘ื–ืžืŸ ืืžืช ืžืื™ื ื˜ืจืืงืฆื™ื” ืขื ื”ืžืฉืชืžืฉ. ืžื—ืฉื‘ื•ืŸ, ื“ืคื“ืคืŸ, ืžื” ืฉืืชื ืจื•ืฆื™ื. ื”ืงื•ื“ ื ื›ืชื‘ ืจืง ื‘ืจื’ืข ืฉื”ื™ื•ื–ืจ ืจื•ืฆื”. ื”ืื ื–ื” ื—ื›ื ื•ื™ืขื™ืœ? ื‘ืจื•ืจ ืฉืœื. ื–ื” ืžื˜ื•ืคืฉ. ื›ืจื’ืข. ืื‘ืœ ืขื ื”ื–ืžืŸ ื”ื“ืžื™ื•ืŸ ื™ื’ื™ืข ื•ื™ื–ืžื™ื ื™ืžืฆืื• ืืช ื”ื™ื•ื–ืงื™ื™ืกื™ื ืฉื‘ื”ื ื–ื” ื™ืขื‘ื•ื“. ืœื›ื, ื‘ืชื•ืจ ื™ื–ืžื™ื ืกื™ื™ื‘ืจ ื•ืชืฉืชื™ื•ืช AI ื™ืฉ ื”ื–ื“ืžื ื•ืช ืœื—ืฉื•ื‘ ืขืœ ื”ืืงื•ืกื™ืกื˜ื ืฉืฆืจื™ืš ืœื‘ื ื•ืช ืกื‘ื™ื‘ ื›ื–ื” ืขื•ืœื. ืื™ืš ืžืื‘ื˜ื—ื™ื ืงื•ื“ ืฉื ื•ืฆืจ ื‘ื–ืžืŸ ืืžืช? ืื™ืš ืžื‘ื˜ื™ื—ื™ื ืจื™ืฆื” ื™ืขื™ืœื”? ืื™ืš ืืคืฉืจ ืœืขืฉื•ืช reuse? ืื’ื‘, ืืคืœื™ืงืฆื™ื•ืช ืฉืžื™ื™ืฆืจื•ืช ืงื•ื“ ื‘ื–ืžืŸ ืืžืช ื•ืคื•ื’ืฉื•ืช ืœืงื•ื—ื•ืช ื›ื‘ืจ ืงื™ื™ืžื•ืช. ืืคื™ืœื• ืœื ื• ื™ืฉ cases ืขื ืžื™ืœื™ื•ื ื™ ืžืฉืชืžืฉื™ื ื—ื•ื“ืฉื™ื™ื. ืื‘ืœ ื›ืžื• ืฉื”ื•ื™ื“ืื• ืžืจืื”, ื–ื” ืจืง ืงืฆื” ื”ืงืจื—ื•ืŸ... ื—ืฉื•ื‘: ื‘ื™ืžื™ื ื”ืืœื” ื™ืฉ ืœื ื• ืงื‘ื•ืฆืช thought leadership ืฉืœ AI ืœื™ื–ืžื™ื, ื›ื“ื™ ืœื“ื‘ืจ ื‘ื“ื™ื•ืง ืขืœ ื”ื“ื‘ืจื™ื ื”ืืœื”, ืชื›ืชื‘ื• ืœื™ ืื ืžืขื ื™ื™ืŸ ืืชื›ื ืœื”ืฆื˜ืจืฃ. ืœื™ื ืง ืœื•ื™ื“ืื• ื‘ืชื’ื•ื‘ื” ื”ื‘ืื”
4
1
1
32
177,548
youtube.com/watch?v=7NfyZhV1โ€ฆ

VibeOS - Fully Hallucinated Operating System

From Microsoft BUILD 2026. vibeOS is by Steve Sanderson

youtube.com
3
565
ื‘ืžืกื’ืจืช ื‘ื™ืช ื”ืœื•ื•ื™ื•ืช ืฉืœ ื”-SaaS, ืกื™ื™ืžื ื• ื‘ื—ื•ื“ืฉ ื”ืื—ืจื•ืŸ ืœืคืชื— ืœืขืฆืžื ื• ืืช ื”ื’ืจืกื” ื”ืจืืฉื•ื ื” ืฉืœ ื”-Notetaker ืฉืœื ื•: ืžืงืœื™ื˜ ืฉื™ื—ื•ืช ืคื ื™ืžื™ ืฉืžืงืœื™ื˜, ืžืชืžืœืœ, ืžืกื›ื ื•ืžืชื—ื‘ืจ ืœื›ืœื™ื ื”ืคื ื™ืžื™ื™ื ืฉืœื ื• ื›ื“ื™ ืœื—ื‘ืจ ืืช ืžื” ืฉืงื•ืจื” ื‘ืฉื™ื—ื” ืœ Agents, ืœ data warehouse ื•ืœืื•ื˜ื•ืžืฆื™ื•ืช ืฉืจืฆื•ืช ืืฆืœื ื• ื‘ื—ื‘ืจื”. ืœืžื” ืœืขื–ืื–ืœ ื—ื‘ืจืช ืฉื™ืจื•ืชื™ AI ืชืจืฆื” ืœืคืชื— ืœืขืฆืžื” Notetaker? ื”ืืžืช ืฉื”ืžื—ื™ืจ ืœื ื”ื™ื” ื”ืฉื™ืงื•ืœ ื”ืขื™ืงืจื™. ืขื 30 ืจื™ืฉื™ื•ื ื•ืช ืฉืขื•ืœื™ื ื›-30 ื“ื•ืœืจ ืœืžื•ืฉื‘, ื”ืชืฉืœื•ื ื”ื•ื ื‘ืขืจืš 900 ื“ื•ืœืจ ื‘ื—ื•ื“ืฉ. ืœื ืžืฉื”ื• ืฉืžืฆื“ื™ืง ืฉื™ื ื•ื™ ืคื•ืงื•ืก. ืื– ืžื” ื›ืŸ? ืงื•ื“ื ื›ื•ืœ, ื‘ืขืœื•ืช ืขืœ ื”-data. ื”ื›ืœื™ื ืฉืขื‘ื“ื ื• ืื™ืชื ืฉืžืจื• ืืช ื”-data ืืฆืœื. ื›ืŸ, ื”ื™ื” ืœื”ื SOC2 ื•-GDPR ื•ื›ื•ืณ, ืื‘ืœ ื›ื•ืœื ื• ืžื‘ื™ื ื™ื ืžื” ื–ื” ืฉื•ื•ื” ื‘ืขื•ืœื ืฉืœ ื”ื™ื•ื. ื”ืžื—ืฉื‘ื” ืฉื”ืฉื™ื—ื•ืช ื”ื›ื™ ืงืจื™ื˜ื™ื•ืช ืฉืœ ื”ื—ื‘ืจื” ืฉืœื ื• ืžืกืชื•ื‘ื‘ื•ืช ืื™ืคืฉื”ื• ื‘ื—ืฉื‘ื•ืŸ ืขื ืŸ ืฉืœ ื—ื‘ืจื” ืฉืœ 200 ืื ืฉื™ื ืฉืคื™ื˜ืจื” ื—ืฆื™ ืฆื•ื•ืช ื‘ื—ื•ื“ืฉ ืฉืขื‘ืจ, ืžื“ืื™ื’ื”. ืื•ืœื™ ื”ื ื™ืžื›ืจื• ืืช ื”-data ืฉืœื ื• ื›ื“ื™ ืœื”ืฆื™ืœ ืืช ืขืฆืžื? ืื•ืœื™ ื”ื ื™ืคื˜ืจื• ืืช ืื™ืฉ ื”-security ื›ื“ื™ ืœื—ืกื•ืš ื‘ืขืœื•ื™ื•ืช? ืžืขื‘ืจ ืœื–ื”, ื”-API ื”ื™ื” ืžื•ื’ื‘ืœ. ื™ืฉ ืœื ื• ื›ืœื™ ืคื ื™ืžื™ ืฉื ื•ืชืŸ ืœืžื ื”ืœื™ื ื“ื•ื—ื•ืช ืขืœ ื”ืชืงื“ืžื•ืช ืฉืœ ืคืจื•ื™ืงื˜ื™ื ืžืชื•ืš ืกื™ื›ื•ืžื™ ืฉื™ื—ื•ืช, ืื‘ืœ ื”-API ื”ื’ื™ืข ืขื ื›ืœ ืžื™ื ื™ ื—ื•ืกืจื™ื: ืื™ ืืคืฉืจ ื”ื™ื” ืœื”ื•ืจื™ื“ ืืช ื”ื•ื•ื™ื“ืื• ื“ืจืš ื”-API, ืื™ ืืคืฉืจ ื”ื™ื” ืœื”ื’ื“ื™ืจ ืžื ื’ื ื•ืŸ retention ืœืคื™ ื—ื•ืงื™ื ื•ื˜ืจื™ื’ืจื™ื ืฉืจืฆื™ื ื•, ื•ืื™ ืืคืฉืจ ื”ื™ื” ืœื ื”ืœ ืืช ื”ืžืฉืชืžืฉื™ื ื‘ืฆื•ืจื” ืฉืžืชืื™ืžื” ืœืืจื’ื•ืŸ ืฉืœื ื•. ืœืžืฉืœ, ืœืืคืฉืจ ืœืžื ื”ืœื™ื ืœืจืื•ืช ืืช ื”ืฉื™ื—ื•ืช ืฉืœ ื›ืœ ื”ืขื•ื‘ื“ื™ื, ื’ื ืื ื”ื ืœื ื—ืœืง ืžืื•ืชื” ืงื‘ื•ืฆื”. ืื– ื›ืฉื”ืชื—ืœื ื• ืœื”ืชืขืกืง ื‘ืื™ื ื˜ื’ืจืฆื™ื•ืช ื”ืืœื”, ืœืžืขืฉื” ื›ื‘ืจ ื”ืชื—ืœื ื• ืœืคืชื— ืืช ื”-SaaS ืžืขืœ ื”ืžืงืœื™ื˜. ื›ืœ ืžื” ืฉื ืฉืืจ ื–ื” ืœืฉืคืจ ืืช ื”ื™ืฆื™ื‘ื•ืช ืฉืœ ื”ืžืงืœื™ื˜ ืขืœ AWS, ืœืขืฉื•ืช ืื•ืคื˜ื™ืžื™ื–ืฆื™ื” ืœืžื›ื•ื ื•ืช ื›ื“ื™ ืœื”ื‘ื˜ื™ื— ื”ืงืœื˜ื” ืื™ื›ื•ืชื™ืช ื•ื™ืฆื™ื‘ื”, ื‘ืืคืจื™ื ื’ ื•ื›ื•ืณ. ื—ื•ื“ืฉ ืฉืœ ืขื‘ื•ื“ื”. ืื—ืจื™ ืฉื”ืขื•ื‘ื“ื™ื ืฉืœื ื• ื”ืชื—ื™ืœื• ืœื”ืฆื˜ืจืฃ ืœืฉื™ื—ื•ืช ื›ืืœื” ืขื ืœืงื•ื—ื•ืช, ืคืชืื•ื ืงื™ื‘ืœื ื• ืฉืืœื•ืช: ืจื’ืข.. ื–ื” ืžืฉื”ื• ืฉืืชื ื™ื›ื•ืœื™ื ืœืขืฉื•ืช ืืฆืœื ื•? ื•ื›ื›ื” ื”ื‘ื ื• ืฉื›ื—ื‘ืจืช ืฉื™ืจื•ืชื™ื ืื ื—ื ื• ื™ื›ื•ืœื™ื ืœื”ืฆื™ืข ืืช ื–ื” ื›ืฉื™ืจื•ืช ืžื ื•ื”ืœ: ืžืงืœื™ื˜ ื–ื•ืœ ื•-self hosted, ืฉื‘ื• ื”-data ื ืฉืžืจ ื‘ืขื ืŸ ืฉืœ ื”ืœืงื•ื—, ื™ื—ื“ ืขื ื›ืœ ื”ืื™ื ื˜ื’ืจืฆื™ื•ืช ืฉื™ื›ื•ืœื•ืช ืœืชืช ืืช ื”-edge ื”ืขืกืงื™ ืฉืžื’ื™ืข ืžื‘ื ื™ื™ืช Agent Workflow ืžืขืœ ืฉื™ื—ื•ืช ืžื•ืงืœื˜ื•ืช. ื•ื”ืจื‘ื” ื™ื•ืชืจ ื–ื•ืœ ืž Gong ืื• ื›ืœื™ื ืื—ืจื™ื.
2
168
Load more