Suhel Mangera
@suhelmangera
Followers
405
Following
642
Media
71
Statuses
1K
Product & data guy | Principal PM ML @ LawVu https://t.co/MuYTSsrEvr
Wellington City, New Zealand
Joined April 2010
My biggest takeaway from @chipro: 1. The biggest improvements to your AI product come from talking to users and understanding their feedback, not from adopting the latest models or staying glued to AI news. Many companies waste time debating which technology to use, when the
What people think improves AI products vs. what actually works with @chipro Chip Huyen was a core developer on @Nvidia’s Nemo platform, a former AI researcher at @Netflix, and taught AI at @Stanford. She’s a two-time founder and the author of two widely read books on AI,
14
34
229
Times, they are a-changin’ “Starting with the next performance review, in addition to asking how much impact each employee delivered we will also ask ourselves how frequently does each person default to Al. If you reach for Google Doc or Sheets before you reach for an Al tool,
42
29
378
Claude Code: no evals [well known code agent company]: no evals [well known code agent company 2]: kinda halfassed evals [leading vibe coding company]: no evals [ceo of company selling you evals]: mmmmm yess all my top customers do evals, you should do evals [vc's in love
Trend I'm following: evals becoming a must-have skill for product builders and AI companies. It's the first new hard skill in a long time that PMs/engineers/founders have had to learn to be successful. The last one was maybe SQL, and Excel? A few examples: @garrytan: "Evals are
130
80
1K
https://t.co/sK0dBjgeu0 Really impressed by the detail in this talk uploaded by @HamelHusain and @sh_reya These were all the lessons I learnt the hard way over the past year but summarised in a one hour talk
2
11
70
So many great quotes from @chipconley on the recent @lennysan's podcast. Here are my 5 favorites: 1. "Culture is what happens when the boss is not around." 2. "Your painful life lessons are the raw material for your future wisdom." 3. "People won't notice your wrinkles as
4
3
19
This 1.5-hour podcast by Sam Altman with Theo Von has a lot of super interesting predictions. - Traditional college will be obsolete for today’s babies. - GPT‑5 already outperforms him on specialized tasks and could manage the full CEO workload “not that long” from now, so
29
76
597
"We finally have proof of an algorithm that can give us expert human reliability and performance, given the right feedback loop. We haven't yet demonstrated long running agentic performance. But you're seeing like the first stumbling steps of that now. And you should see much
15
13
197
I wonder if these investigations can be sped by leveraging AI and LLM to comb through documents, videos and audio. It seems baffling that it takes 6 years to investigate one case! @rnz_news @radionz
https://t.co/8ifj4JYXCb
rnz.co.nz
A Bangladeshi couple have been found guilty of 40 charges of immigration and identity fraud spanning two decades.
0
0
0
Introducing the Responses API: the new primitive of the OpenAI API. It is the culmination of 2 years of learnings designing the OpenAI API, and the foundation of our next chapter of building agents. 🧵Here’s the story of how we designed it:
117
364
3K
Agency > Intelligence I had this intuitively wrong for decades, I think due to a pervasive cultural veneration of intelligence, various entertainment/media, obsession with IQ etc. Agency is significantly more powerful and significantly more scarce. Are you hiring for agency? Are
1K
7K
37K
@Meta @zynga @Headspace @VRChat Some key takeaways: 1. Product strategy sits between mission/vision and execution plans, forcing choices about where to deploy scarce resources for maximum impact. A good strategy includes three key components: a. A handful of focus areas (strategic pillars) b. Areas explicitly
0
2
25
World's First Black Pilot in (1916), Ottoman Empire Pilot Ahmet Ali Effendi.
19
84
707
Among all of the open source AI labs, DeepSeek has perhaps made the most architectural improvements to the standard decoder-only Transformer. In this week’s Gradient Updates issue, @EgeErdil2 breaks down the key architectural innovations behind DeepSeek’s latest model. 🧵
12
35
363
Where is AI going? Six leaders share their hopes for AI in the coming year, in The Batch: - Hanno Basse: Generative AI for Artists - David Ding: Generated Video With Music, Sound Effects, and Dialogue - Joseph Gonzalez: General Intelligence - Albert Gu: More Learning, Less Data -
55
220
1K
Web scraping will never be the same! Crawl4AI simplifies web crawling and data extraction, making it ready to use for LLMs and AI applications. Here’s why it’s a game-changer: 🆓 Completely free and open-source 🚀 Blazing fast performance, outperforming many paid services 🤖
37
332
2K
This is the most important paper in a long time . It shows with strong evidence we are reaching the limits of quantization. The paper says this: the more tokens you train on, the more precision you need. This has broad implications for the entire field and the future of GPUs🧵
[1/7] New paper alert! Heard about the BitNet hype or that Llama-3 is harder to quantize? Our new work studies both! We formulate scaling laws for precision, across both pre and post-training https://t.co/8FC4g2fSRb. TLDR; - Models become harder to post-train quantize as they
64
480
3K
woah! just noticed @readwise and NotebookLM have an integration. You can export and keep in sync and chat with all your highlights or generate podcast. going to try it out and post notes later on https://t.co/g6mtk0lkKp
0
0
0
why do ppl have to fold their arms in pics? seems so unnatural
0
0
0
I’m excited to announce a partnership with @Perplexity_AI. Starting today, all Lenny's Newsletter paid subscribers will get a free year of Perplexity Pro. For just $150/year, a paid subscription to my newsletter now includes Perplexity Pro (a $240/year value), an invite to a
62
47
1K