HackerLangs
TopNewTrendsCommentsPastAskShowJobs

joozio

629 karmajoined il y a 8 mois
Just a curious tech enthusiast sharing digital experiments and learnings along the way. Writing about AI, e-commerce, and the future of work - mistakes and discoveries included.

Submissions

The Download: Claude's inner workings and OpenAI's "super app"

technologyreview.com
1 points·by joozio·il y a 10 heures·0 comments

The Download: a nuclear landmark, and China eyes Nvidia chips

technologyreview.com
1 points·by joozio·hier·0 comments

Four nuclear reactors hit a big milestone in the US

technologyreview.com
11 points·by joozio·avant-hier·0 comments

Google pays $250K for Linux vulnerability allowing guest VM escapes

arstechnica.com
3 points·by joozio·avant-hier·0 comments

EmTech AI 2026: The Rise of the AI Platform

technologyreview.com
1 points·by joozio·avant-hier·0 comments

Open-source LLMs administer maximum electric shocks in a Milgram-like obedience

lesswrong.com
1 points·by joozio·il y a 3 jours·0 comments

Hackers can use 9 of the most popular AI tools to assemble botnets

arstechnica.com
5 points·by joozio·il y a 3 jours·0 comments

The Download: your stake in OpenAI, and the Treasury's AI warning

technologyreview.com
1 points·by joozio·il y a 3 jours·0 comments

The foundational elements of AI architecture that IT leaders need to scale

technologyreview.com
2 points·by joozio·il y a 3 jours·0 comments

Why worms (and microbes) are catching on as a manure pollution solution

technologyreview.com
2 points·by joozio·il y a 4 jours·0 comments

Family's $300 Stake in OpenAI

technologyreview.com
3 points·by joozio·il y a 4 jours·0 comments

The Download: South Korea's hottest bachelors, and advancing eye transplants

technologyreview.com
2 points·by joozio·il y a 4 jours·0 comments

South Korea's hottest new bachelors are chip workers

technologyreview.com
19 points·by joozio·il y a 5 jours·3 comments

When Gemma Thinks About Resources – It Fails: A Behavioral Experiment

lesswrong.com
3 points·by joozio·il y a 5 jours·0 comments

Probing the loss-band sparsity assumption in Scientist AI

lesswrong.com
2 points·by joozio·il y a 5 jours·0 comments

All Your Favorite Gadgets Are Getting More Expensive Again

wired.com
5 points·by joozio·il y a 6 jours·0 comments

Teaching AI to Run with the Turbines

technologyreview.com
1 points·by joozio·il y a 7 jours·0 comments

A device that revives eyeballs from dead donors could make eye transplants poss

technologyreview.com
10 points·by joozio·il y a 7 jours·0 comments

The Download: a smoking "endgame" and a new Elizabeth Bear story

technologyreview.com
2 points·by joozio·il y a 7 jours·0 comments

The UK's generational tobacco ban might not work. I'm supporting it anyway

technologyreview.com
2 points·by joozio·il y a 8 jours·3 comments

comments

joozio
·il y a 17 jours·discuss
Meh. They are still behind current Agents mainstream IMHO
joozio
·il y a 24 jours·discuss
[dead]
joozio
·il y a 2 mois·discuss
[dead]
joozio
·il y a 3 mois·discuss
Funny. I thought I was the only one. Then I found more people and now you wrote about that. Just this week I also wrote about Claude Opus 4.7 and how I came back to Codex after that: https://thoughts.jock.pl/p/opus-4-7-codex-comeback-2026
joozio
·il y a 3 mois·discuss
Terminal Benchmark: https://www.tbench.ai/leaderboard/terminal-bench/2.0
joozio
·il y a 3 mois·discuss
I run a Claude Code agent 24/7 on a Mac Mini. After a few months my morning routine was gone and I was reviewing agent output at midnight. Built this to teach it boundaries.

The interesting part ended up being the error registry. Agents fail silently way more than you'd expect. Same error repeats 50 times burning tokens before you notice.

Zero dependencies, Python stdlib only. Would love feedback on what's missing.
joozio
·il y a 4 mois·discuss
I think not yet, but Anthropic is trying to.
joozio
·il y a 4 mois·discuss
I am user of Arc(even now!) and I really love it. Feels different in a good way. I can see it's PoC, but have you touched backend even in concept?
joozio
·il y a 4 mois·discuss
Seperate space, more power and also Local small LLMs. Also 24/7 :)
joozio
·il y a 4 mois·discuss
[dead]
joozio
·il y a 4 mois·discuss
Anthropic formalized their enterprise partnership program today. Key partners include Accenture (30k professionals trained on Claude), Cognizant (350k associates with Claude access), Deloitte, and Infosys.

What makes this interesting: Claude is the only frontier model available across all three major cloud platforms simultaneously (AWS Bedrock, Azure, Google Cloud). The Partner Network is how they convert that platform coverage into an actual distribution advantage.

The $100M goes toward: certification programs (Claude Certified Architect launched today), dedicated Applied AI engineers for customer engagements, sales playbooks, and co-marketing.

This signals a shift from "try our model" to "we'll help you deploy org-wide." Curious if others are seeing this pattern play out in their enterprise conversations -- is this differentiated from what OpenAI/Google are doing, or is the whole industry moving this direction?
joozio
·il y a 4 mois·discuss
Nope - written by me.
joozio
·il y a 5 mois·discuss
Thanks! The investment angle is interesting — I hadn't thought about it that way, but it makes sense. If you're seeing the gap firsthand, you have an information edge most investors don't.

What strikes me most is how different the conversation is depending on where you are. Reddit investment subs, Twitter AI circles, and actual workplaces — three completely different realities about the same technology.

I think the key thing that's hard to convey to non-users is the compounding effect. Once you hit a certain depth, every new tool or workflow multiplies what you already know. My neighbor who codes with Gemini is one "aha moment" away from a completely different relationship with AI — but that moment hasn't happened yet for most people.

The gap you're betting on seems real to me. Whether it closes in months or years is the interesting question.
joozio
·il y a 5 mois·discuss
Love this idea. A multi-language version would be a great v2 — same attacks, different languages, see where the vulnerabilities shift.
joozio
·il y a 5 mois·discuss
Haven't benchmarked pre-processing approaches yet, but that's a natural next step. Right now the test page targets raw agent behavior — no middleware. A comparison between raw vs sanitized pipelines against the same attacks would be really useful. The multi-layer attack (#10) would probably be the hardest to strip cleanly since it combines structural hiding with social engineering in the visible text.
joozio
·il y a 5 mois·discuss
It's working -> your agents scored A+, which means they resisted all 10 injection attempts. That's a great result. The tool detects when canary phrases leak into the response. If nothing leaked, you get a clean score. Not all models are this resilient though - we've seen results ranging from A+ to C depending on the model and even the language used.
joozio
·il y a 5 mois·discuss
That's a really interesting edge case - screenshot-based agents sidestep the entire attack surface because they never process raw HTML. All 10 attacks here are text/DOM-level. A visual-only agent would need a completely different attack vector (like rendered misleading text or optical tricks). Might be worth exploring as a v2.
joozio
·il y a 5 mois·discuss
Great point -> just shipped an update based on this. The tool now distinguishes three states: Resisted (ignored it), Detected (mentioned it while analyzing/warning), and Compromised(actually followed the instruction). Agents that catch the injections get credit for detection now.
joozio
·il y a 5 mois·discuss
The idea, design, and decisions were mine. I use Claude Code as a dev tool, same as anyone using Copilot or Cursor. The 'night shift' framing was maybe bad fit here.
joozio
·il y a 5 mois·discuss
I never thought that multi-language could be a factor here...