Hacker News: harisec

New comment by harisec in "AI is wiping out entry-level tech jobs, leaving graduates stranded"

harisec — Tue, 16 Dec 2025 20:17:59 +0000

Unfortunatelly, i have the same experience.

Security research in the age of AI tools

harisec — Wed, 03 Dec 2025 14:36:09 +0000

Article URL: https://www.invicti.com/blog/security-labs/security-research-in-the-age-of-ai-tools

Comments URL: https://news.ycombinator.com/item?id=46134943

Points: 2

# Comments: 0

New comment by harisec in "CVE program faces swift end after DHS fails to renew contract"

harisec — Wed, 16 Apr 2025 10:25:06 +0000

$44M/year?

https://www.usaspending.gov/award/CONT_AWD_70RCSJ23FR0000015...

New comment by harisec in "The Llama 4 herd"

harisec — Sun, 06 Apr 2025 17:04:56 +0000

Yes, during training multiple checkpoints are created, you can distill from any checkpoint you want.

My Thoughts on the Future of "AI"

harisec — Fri, 14 Mar 2025 08:32:42 +0000

Article URL: https://nicholas.carlini.com/writing/2025/thoughts-on-future-ai.html

Comments URL: https://news.ycombinator.com/item?id=43360627

Points: 2

# Comments: 0

New comment by harisec in "Grok3 Launch [video]"

harisec — Tue, 18 Feb 2025 13:06:38 +0000

Anybody can try Grok3 on Chatbot Arena (even if you are in Europe). Select Direct Chat and select the model early-grok-3. https://lmarena.ai/

Coinbase, Binance Plan to List Donald Trump's Official Trump Token

harisec — Sun, 19 Jan 2025 17:50:57 +0000

Article URL: https://www.coindesk.com/markets/2025/01/19/coinbase-binance-plan-to-list-donald-trumps-official-trump-token-after-its-phenomenal-debut

Comments URL: https://news.ycombinator.com/item?id=42759421

Points: 2

# Comments: 0

How to jailbreak most/all LLMs using Assistant Prefill

harisec — Fri, 10 Jan 2025 15:28:34 +0000

Article URL: https://www.invicti.com/blog/security-labs/first-tokens-the-achilles-heel-of-llms/

Comments URL: https://news.ycombinator.com/item?id=42656466

Points: 3

# Comments: 0

New comment by harisec in "Bluesky is currently gaining more than 1M users a day"

harisec — Sun, 17 Nov 2024 11:19:55 +0000

It doesn’t matter much how many users Bluesky is gaining, it matters how many of them will be using Bluesky in a few months. We will see.

New comment by harisec in "Francois Chollet is leaving Google"

harisec — Thu, 14 Nov 2024 13:51:40 +0000

Congrats, good luck with your new company!

I have one question regarding your ARC Prize competition: The current leader from the leaderboard (MindsAI) seems not to be following the original intention of the competition (fine tune a model with millions of tasks similar with the ARC tasks). IMO this is against the goal/intention of the competition, the goal being to find a novel way to get neural networks to generalize from a few samples. You can solve almost anything by brute-forcing it (fine tunning on millions of samples). If you agree with me, why is the MindsAI solution accepted?

New comment by harisec in "System prompt exposure: how AI image generators may leak sensitive instructions"

harisec — Tue, 12 Nov 2024 13:03:20 +0000

Recraft’s image generation service could leak its internal system prompts due to its unique architecture combining Claude (an AI language model) with a diffusion model. Unlike other image generators, Recraft could perform calculations and answer questions, which led to the discovery that carefully crafted prompts could expose the system’s internal instructions.

System prompt exposure: how AI image generators may leak sensitive instructions

harisec — Tue, 12 Nov 2024 13:03:20 +0000

Article URL: https://www.invicti.com/blog/security-labs/system-prompt-exposure-how-ai-image-generators-may-leak-sensitive-instructions/

Comments URL: https://news.ycombinator.com/item?id=42115162

Points: 1

# Comments: 1

Analyzing WordPress hack access logs with NotebookLM

harisec — Tue, 29 Oct 2024 09:59:33 +0000

Article URL: https://www.invicti.com/blog/security-labs/analyzing-wordpress-hack-access-logs-with-notebooklm/

Comments URL: https://news.ycombinator.com/item?id=41981818

Points: 1

# Comments: 0

New comment by harisec in "Everything I built with Claude Artifacts this week"

harisec — Thu, 24 Oct 2024 12:20:13 +0000

Actually, i think it will take less than 2 years. I've been using Aider + Claude 3.5 Sonnet almost daily for a long time and the progress is very fast. We will see.

New comment by harisec in "Everything I built with Claude Artifacts this week"

harisec — Thu, 24 Oct 2024 06:24:53 +0000

That's a good point and I agree with you. However, would you agree that in a few years we will need far less developers than we need right now?

New comment by harisec in "Everything I built with Claude Artifacts this week"

harisec — Thu, 24 Oct 2024 06:08:09 +0000

These are toys but in 2 years they will probably be full projects and 2 years later people will ask "why do i need a software developer?"

New comment by harisec in "Everything I built with Claude Artifacts this week"

harisec — Thu, 24 Oct 2024 06:04:36 +0000

If you want to really get depressed about the future of software developers try aider.

New comment by harisec in "Un Ministral, Des Ministraux"

harisec — Wed, 16 Oct 2024 16:16:14 +0000

I never had this problem but i guess it depends on the prompt.

New comment by harisec in "Un Ministral, Des Ministraux"

harisec — Wed, 16 Oct 2024 15:46:07 +0000

Qwen 2.5 models are better than Llama and Mistral.

New comment by harisec in "Famous AI Artist Says He's Losing Millions from People Stealing His Work"

harisec — Thu, 03 Oct 2024 04:47:02 +0000

There are no "LLM-generated images". These images are generated using diffusion models. LLMs are very different from diffusion models.

https://en.m.wikipedia.org/wiki/Diffusion_model