Hacker News: fsndz

New comment by fsndz in "The Learning Loop and LLMs"

fsndz — Thu, 06 Nov 2025 23:12:58 +0000

we can't automate it anyway and vibe coding is overrated: https://medium.com/thoughts-on-machine-learning/vibe-coding-...

New comment by fsndz in "You should write an agent"

fsndz — Thu, 06 Nov 2025 23:09:35 +0000

I did that, burned 2.6B tokens in the process and learned a lot: https://transitions.substack.com/p/what-burning-26-billion-p...

New comment by fsndz in "[dead]"

fsndz — Wed, 29 Oct 2025 21:11:00 +0000

As part of the Holistic Agent Leaderboard (HAL) initiative at Princeton CITP, we evaluated more than 220 agent runs across 9 benchmarks, the equivalent of over 20,000 agent rollouts across 9 models and 9 benchmarks for a total cost of $40,000. The benchmarks are: AssistantBench, CORE-Bench Hard, GAIA, Online Mind2Web, Scicode, ScienceAgentBench, SWE-bench Verified Mini, TAU-bench Airline, and USACO.

In that process, we “burned” 2.6 billion prompt tokens and learned a lot along the way. In this article, I’d like to share some of the insights we gained, with a particular focus on the GAIA benchmark.

New comment by fsndz in "ChatGPT agent: bridging research and action"

fsndz — Thu, 17 Jul 2025 22:20:14 +0000

By that definition, the ChatGPT app is now an AI agent. When you use ChatGPT nowadays, you can select different models and complement these models with tools like web search and image creation. It’s no longer a simple text-in / text-out interface. It looks like it is still that, but deep down, it is something new: it is agentic… https://medium.com/thoughts-on-machine-learning/building-ai-...

New comment by fsndz in "Cognition (Devin AI) to Acquire Windsurf"

fsndz — Mon, 14 Jul 2025 20:52:27 +0000

and people are still saying vibe coding is overrated? nonsense: https://www.lycee.ai/blog/why-vibe-coding-is-overrated

New comment by fsndz in "Cognition (Devin AI) to Acquire Windsurf"

fsndz — Mon, 14 Jul 2025 20:50:26 +0000

Exactly. I think the study is a good reminder that we really have to be careful about the productivity gains attributed to AI. Main takeaway imo, despite limitations from the study, is AI is not a panacea, it can increase productivity, but only if used 'well' and with the good workflows in place, and in the right context.

New comment by fsndz in "Cognition (Devin AI) to Acquire Windsurf"

fsndz — Mon, 14 Jul 2025 20:49:46 +0000

I mean, hacker news is still the same aren't they using AI to completely make this website more of whatever it was before ????

New comment by fsndz in "Amazon CEO says AI agents will soon reduce company's corporate workforce"

fsndz — Sun, 13 Jul 2025 21:53:47 +0000

exactly !

New comment by fsndz in "Amazon CEO says AI agents will soon reduce company's corporate workforce"

fsndz — Sun, 13 Jul 2025 19:58:37 +0000

Klarna tried this narrative shopping strategy first and it backfired: https://fsndzomga.medium.com/i-have-no-confidence-in-klarna-...

New comment by fsndz in "Amazon CEO says AI agents will soon reduce company's corporate workforce"

fsndz — Sun, 13 Jul 2025 19:55:58 +0000

I am tired of seeing this. First the Klarna dude did it and it backfired. Now Andy... People fail to grasp the fact that building AI agents is no longer enough, you need to do more: https://medium.com/thoughts-on-machine-learning/building-ai-...

New comment by fsndz in "Supabase MCP can leak your entire SQL database"

fsndz — Tue, 08 Jul 2025 22:17:40 +0000

I now understand why some people say MCP is mostly bullshit + a huge security risk: https://www.lycee.ai/blog/why-mcp-is-mostly-bullshit

New comment by fsndz in "Claude 4"

fsndz — Thu, 22 May 2025 20:08:07 +0000

It's not. can we stop doing this ?

New comment by fsndz in "Getting AI to write good SQL: Text-to-SQL techniques explained"

fsndz — Fri, 16 May 2025 22:15:47 +0000

the smolagents library is also pretty nice to do the scaffolding around the model. Text to sql seems simple in demos, but to make it work in real life complex cases is very hard: https://medium.com/thoughts-on-machine-learning/build-a-text...

New comment by fsndz in "I don't like NumPy"

fsndz — Thu, 15 May 2025 21:23:37 +0000

xarray is nice

New comment by fsndz in "The unreasonable effectiveness of an LLM agent loop with tool use"

fsndz — Thu, 15 May 2025 21:11:48 +0000

I can be frustrating at times. but my experience is the more you try the better you become at knowing what to ask and to expect. But I guess you understand now why some people say vibe coding is a bit overrated: https://www.lycee.ai/blog/why-vibe-coding-is-overrated

New comment by fsndz in "Baby is healed with first personalized gene-editing treatment"

fsndz — Thu, 15 May 2025 21:07:48 +0000

Never bet against science !

New comment by fsndz in "Evolving OpenAI's Structure"

fsndz — Mon, 05 May 2025 19:39:17 +0000

it's always opt-in until it isn't

New comment by fsndz in "Supabase raises $200M Series D at $2B valuation"

fsndz — Tue, 22 Apr 2025 19:11:01 +0000

and setting up postgresql on a simple VPS is so easy... You can literally ask Gemini 2.5 Pro or o3 or Sonnet 3.7 and do it in 15-30 minutes... Learned helplessness is really something and vibe coding is overrated imo: https://www.lycee.ai/blog/why-vibe-coding-is-overrated

New comment by fsndz in "Gemini 2.5 Flash"

fsndz — Fri, 18 Apr 2025 20:31:24 +0000

why is Google more evil than say OpenAI ?

New comment by fsndz in "Gemini 2.5 Flash"

fsndz — Thu, 17 Apr 2025 22:22:25 +0000

More and more people are coming to the realisation that Google is actually winning at the model level right now.