@codewithjerry01

AI tips and news without the hype, plus plain-English breakdowns of new models.

 Nigeria
tiktok.com/@codewithjerry01?_r=1&_t=ZN-96ubPr1jCsA
Joined December 2025
  • Tweets2.7K
  • Following643
  • Followers944
280Media
Hi, I'm Jerry, a self-taught developer from Nigeria. I post plain-English AI news and tips, no hype, and I build websites and AI automations for local businesses. Follow if you want AI explained simply. What should I break down next? 👇
3
5
16
959
Claude Duet (research preview) works in the background and only pings you when something needs attention. Would you let it run part of your workday? Yes Small tasks only Not yet
2
- Codex - Claude - Gemini @grok Which model is this ?
2
37
Would you let AI run your inbox and calendar for a week? > Yes > Only drafts > Never
22
Flint runs on RTX Spark and DGX Spark systems, GeForce RTX and RTX PRO GPUs, and M4 Pro. Model weights are available on HuggingFace! Test it out 👀 huggingface.co/RecursiveSI/f…

RecursiveSI/flint · Hugging Face

We’re on a journey to advance and democratize artificial intelligence through open source and open science.

huggingface.co
9
Anthropic could reportedly make its S-1 IPO filing public as early as Friday. Pre-IPO traders on Liquid are pricing it near $2 trillion, but that's market speculation, not an official valuation. Too high, too low, or fair?
Liquid
@liquidtrading18h
JUST IN: Anthropic is set to publicly file its S-1 investor prospectus as soon as this Friday, according to three people familiar with the artificial intelligence firm's plans.
1
1
42
Me when the Boss is trying to get someone to work Overtime. Why always me?😩😂
1
1
63
Are we really cooked? 2022 - Student 2023 - Developer 2024 - Prompt engineer 2025 - Vibe coder 2026 - Al agent 2027 - Farmer
2
2
5
135
Be honest: How true is this?
4
118
Code With Jerry | AI Tips retweeted
philip.jpg
@pelloiafilippoOct 20
Hey guys, don't give up.
40
4
2
109
1,885
Me and Claude everyday
41
Code With Jerry | AI Tips retweeted
AM09:21
@AM09_21Oct 20
DeepSeek、あのとち狂った料金はほぼ赤字なんじゃないかな〜とか思ってたけど、公開された推論カーネルであるDeepinfraから逆算して計算したら原価がアホほど安かったんよね 粗利率は80%〜95%くらい……() GPTとかClaudeのAPI料金って目ん玉飛び出るほど高いけど、原価はどうなんだろうね とはいえ今のAIに割かれてる予算はたいていモデルの訓練段階だし、それも込みで考えたらDeepSeekの価格設定は本当にぴったりって感じ ここまで大人気なモデルになってるなら、元取っててもおかしくないね
27
90
11
995
102,027
I think we should take a break. Like WTF
1
1
96
If one got a body tomorrow, what would its first move be?
38
Alexandr Wang explains how Al is made in 3 core components.
37
For real, if they keep improving this model, future AI releases will have a serious competitor.
Theo - t3.gg
@theoOct 20
Opus 5.5 is proof that the world does actually need smarter models.
29
Eagle released EAGLE-4, an open-source drafter for speeding up LLM inference. By its own benchmark, Qwen3.8-27B hit 961.6 tokens/s on 4x NVIDIA B300, about 4x vanilla. Vendor numbers, so test it yourself. Would you switch your inference stack?
Eagle
@EagleCorpOct 20
EAGLE-4 is here. From EAGLE-1 to EAGLE-3, we’ve kept making inference faster. EAGLE-4 is our fastest yet. 961.6 tokens/s on @Alibaba_Qwen’s Qwen3.8-27B with TokenSpeed (BF16, 4× @NVIDIA B300, TP4, one request, T=0). Github: github.com/eagle-hq/EAGLE-4
1
55
Daily AI Brief, Oct 20: • ChatGPT Slides: real-time co-editing • Anthropic reportedly ships a new model this month • Meta's Muse Glimmer: 30B open model, one consumer GPU • Inception's Mercury 3: 1,000+ tokens/sec Which one matters most?
1
2
51
LEARN ANY SKILL FROM ZERO USING CLAUDE AS YOUR PERSONAL TUTOR🥈👇👇
1
1
45
Get access to the prompts👇👇 docs.google.com/document/d/1…
27
Developers, what's the first thing you open every morning? X Your code AI chat Email
2
1
75
I open X first thing in the morning
30
Load more