Harsh Varagiya

1.3K posts

Harsh Varagiya

Harsh Varagiya

@HarshVaragiya

cybersecurity professional

127.0.0.1 Katılım Ekim 2017
3.3K Takip Edilen307 Takipçiler
Harsh Varagiya retweetledi
Harald Schäfer
Harald Schäfer@___Harald___·
To know what Chinese labs are doing you can just read their papers. To know what American labs are doing you have to wait for @karpathy to do a 1 year internship and then post a GitHub repo with what he learned.
Andrej Karpathy@karpathy

Personal update: I've joined Anthropic. I think the next few years at the frontier of LLMs will be especially formative. I am very excited to join the team here and get back to R&D. I remain deeply passionate about education and plan to resume my work on it in time.

English
26
157
3.9K
149.2K
Harsh Varagiya retweetledi
chompie
chompie@chompie1337·
Claude helped me with this bug too but in a different way... Tried to gaslight me saying it wasn’t ~exploitable in practice~ and I got obsessed with proving it wrong 😩
TrendAI Zero Day Initiative@thezdi

Confirmed! @chompie1337 of IBM X-Force Offensive Research (XOR) used a race condition to escalate privileges on Red Hat Enterprise Linux for Workstations, earning $20,000 and 2 Master of Pwn points. #Pwn2Own #P2OBerlin

English
42
102
1.3K
74.9K
Harsh Varagiya retweetledi
Harsh Varagiya retweetledi
ThePrimeagen
ThePrimeagen@ThePrimeagen·
There is a certain side of running a business that these AI hypes are missing. Always running at capacity leaves no room for really thinking about what is important Sure the AI is "letting you do more than ever," but those same people stopped dreaming and are always executing. Feels like a tragedy
English
146
85
1.6K
61.1K
Harsh Varagiya retweetledi
ThePrimeagen
ThePrimeagen@ThePrimeagen·
This
ThePrimeagen tweet media
English
234
948
12.8K
423K
Harsh Varagiya retweetledi
Sudo su
Sudo su@sudoingX·
hot take: 90% of ai startups paying for api calls could run the same workloads locally on a single 3090 and never notice the difference. you don't need frontier pricing for tasks a 27B model handles fine. most have never even tested a quantized model on consumer hardware. not every task in your pipeline should be burning credits. audit your workload. you'd be surprised what runs locally.
English
71
23
430
44.9K
Harsh Varagiya retweetledi
solst/ICE of Astarte
solst/ICE of Astarte@IceSolst·
Cybersecurity is just software engineering’s PvP mode
English
55
664
6.4K
167.1K
Harsh Varagiya retweetledi
ib
ib@Indian_Bronson·
“We looked at 1M conversations” Surprise!
English
38
653
17.6K
1.3M
Harsh Varagiya retweetledi
Bojie Li
Bojie Li@bojie_li·
Closed labs hide model sizes. They can't hide what their models know, and what a model knows is an indicator on how big it is. Reasoning compresses. Factual knowledge doesn't. So you can size a frontier model from black-box API calls alone, and across releases you can literally watch a single fact arrive in the parameters over time. For three years, my friends Jiyan He and Zihan Zheng have been asking frontier LLMs the same question: "what do you know about USTC Hackergame?", a CTF contest. May 2024: GPT-4o invented fake titles. Feb 2025: Claude 3.7 Sonnet listed 19 verified 2023 challenges. By April 2026, frontier models recall specific challenges across consecutive years. After DeepSeek-V4 dropped, I instructed my agent to spend four days autonomously turning that habit into Incompressible Knowledge Probes (IKP) — 1,400 questions, 7 tiers of obscurity, 188 models, 27 vendors. Three findings: 1/ You can approximately size any black-box LLM from factual accuracy alone. Penalized accuracy is log-linear in log(params), R² = 0.917 on 89 open-weight models from 135M to 1.6T params. Project closed APIs onto the curve → GPT-5.5 ~9T, Claude Opus 4.7 ~4T, GPT-5.4 ~2.2T, Claude Sonnet 4.6 ~1.7T, Gemini 2.5 Pro ~1.2T (90% CI: 0.3-3x size). 2/ Citation count and h-index don't predict whether a frontier model recognizes a researcher. Two researchers with similar citation profiles get very different responses. Models memorize impact — work that shaped a field, not many incremental papers. 3/ Factual capacity doesn't compress over time. Across 96 open-weight models across 3 years, the IKP time coefficient is statistically zero, rejecting the Densing-Law prediction of +0.0117/month at p<10⁻¹⁵. Reasoning benchmarks saturate; factual capacity keeps scaling with parameters. Website: 01.me/research/ikp/ Paper: arxiv.org/pdf/2604.24827
Bojie Li tweet mediaBojie Li tweet mediaBojie Li tweet media
English
71
233
2.2K
388.7K
Harsh Varagiya retweetledi
Bo Wang
Bo Wang@BoWang87·
Totally agree that AlphaFold didn't “solve” protein folding! A system that accurately predicts final structures hasn't explained why those residues fold that way, eg., the energy landscape, the kinetic pathways, what happens co-translationally before the chain is even released from the ribosome… etc “Solving” means understanding the mechanism. That's a different kind of question. It's the difference between predicting where a ball lands and understanding gravity. Without that, we can't explain misfolding diseases from first principles, design truly novel protein architectures, or predict how mutations shift folding kinetics rather than just final structure. AlphaFold gave us better maps. The physics of folding is still largely uncharted.
Dr Alexander D. Kalian@AlexanderKalian

Every time I tell AI utopianists that biology is too complex for AI to "solve", they cite the success of AlphaFold. No, AlphaFold did not "solve" protein folding. It gets broad structures correct ~70-88% of the time (depending on evaluation), enabling useful but flawed statistical guesses. True "solving" would require ~99.9%+ accuracy, practically zero meaningful edge cases, and high confidence across fine details like side chains and conformations. Even then, this is just one narrow slice of the complexities of proteomics. The persistent gap between the "AlphaFold solved protein folding" claim and reality is a perfect example of AI overhype in biology.

English
16
39
320
78.6K
Harsh Varagiya retweetledi
Deli Chen
Deli Chen@victor207755822·
DeepSeek-V3: Dec 26, 2024 DeepSeek-V4: Apr 24, 2026 484 days later, we humbly share our labor of love. As always, we stay true to long-termism and open source for all. AGI belongs to everyone. ❤️🌍 #DeepSeekV4 #AGIforEveryone #OpenSource
DeepSeek@deepseek_ai

🚀 DeepSeek-V4 Preview is officially live & open-sourced! Welcome to the era of cost-effective 1M context length. 🔹 DeepSeek-V4-Pro: 1.6T total / 49B active params. Performance rivaling the world's top closed-source models. 🔹 DeepSeek-V4-Flash: 284B total / 13B active params. Your fast, efficient, and economical choice. Try it now at chat.deepseek.com via Expert Mode / Instant Mode. API is updated & available today! 📄 Tech Report: huggingface.co/deepseek-ai/De… 🤗 Open Weights: huggingface.co/collections/de… 1/n

English
353
1.3K
13.1K
1M
Harsh Varagiya retweetledi
DeepSeek
DeepSeek@deepseek_ai·
🚀 DeepSeek-V4 Preview is officially live & open-sourced! Welcome to the era of cost-effective 1M context length. 🔹 DeepSeek-V4-Pro: 1.6T total / 49B active params. Performance rivaling the world's top closed-source models. 🔹 DeepSeek-V4-Flash: 284B total / 13B active params. Your fast, efficient, and economical choice. Try it now at chat.deepseek.com via Expert Mode / Instant Mode. API is updated & available today! 📄 Tech Report: huggingface.co/deepseek-ai/De… 🤗 Open Weights: huggingface.co/collections/de… 1/n
DeepSeek tweet media
English
1.6K
7.7K
45.4K
9.7M
Harsh Varagiya retweetledi
Strace
Strace@straceX·
this is the most satisfying plot twist in tech history
Strace tweet media
English
213
8.8K
110.6K
1.6M
Harsh Varagiya
Harsh Varagiya@HarshVaragiya·
@Fossil customer support is refusing to share an invoice copy of my purchase and also refusing to honor the warranty on my watch (despite registering it on their site). I understand losing an invoice due to some billing platform issue, but why should customers suffer for that ?
English
0
0
0
25
NVIDIA GeForce
NVIDIA GeForce@NVIDIAGeForce·
PRAGMATA has launched with #RTXON, featuring path tracing and DLSS 4! To celebrate, we are giving away this custom wrapped GeForce RTX 5090 featuring Hugh and Diana, perfect for the adventure that awaits on the moon. Want it? Comment "PRAGMATA RTX" to enter!
NVIDIA GeForce tweet media
English
30.1K
3.1K
20.7K
1.7M