Heiko Hotz
@heikohotz.bsky.social
AI Engineer @ Google 👨💻 — Educator 👨🏫 — Traveller ✈️ — Hobby photographer 📷 — Foodie 🌮 — Film fan 🍿 — Boardgamer 🎲 — Londoner💂♂️
Medium: https://heiko-hotz.medium.com/
Github: https://github.com/heiko-hotz
LI: https://www.linkedin.com/in/heikohotz/
Medium: https://heiko-hotz.medium.com/
Github: https://github.com/heiko-hotz
LI: https://www.linkedin.com/in/heikohotz/
Oh, hello! May I meet you?
November 18, 2025 at 3:58 PM
Oh, hello! May I meet you?
Official results are in - Gemini achieved gold-medal level in the International Mathematical Olympiad! 🏆
July 21, 2025 at 10:22 PM
Official results are in - Gemini achieved gold-medal level in the International Mathematical Olympiad! 🏆
Evaluating voice-driven agents got you pulling your hair out? 😩 Evaluating voice agents is WILD. Accents, noise, weird speech... how do you even test?! Manual prompt engineering for that? A total nightmare. 👇
July 15, 2025 at 7:05 AM
Evaluating voice-driven agents got you pulling your hair out? 😩 Evaluating voice agents is WILD. Accents, noise, weird speech... how do you even test?! Manual prompt engineering for that? A total nightmare. 👇
Introducing Gemini-Powered Slide Creation by Voice!
In this quick demo, I’ve integrated a “Slide Creation Agent” into my personal project, Project Pastra. Watch how it effortlessly generates slides based on voice instructions.
In this quick demo, I’ve integrated a “Slide Creation Agent” into my personal project, Project Pastra. Watch how it effortlessly generates slides based on voice instructions.
January 14, 2025 at 7:47 AM
Introducing Gemini-Powered Slide Creation by Voice!
In this quick demo, I’ve integrated a “Slide Creation Agent” into my personal project, Project Pastra. Watch how it effortlessly generates slides based on voice instructions.
In this quick demo, I’ve integrated a “Slide Creation Agent” into my personal project, Project Pastra. Watch how it effortlessly generates slides based on voice instructions.
The Gemini Multimodal Live API Developer Guide is live!
December 30, 2024 at 9:25 AM
The Gemini Multimodal Live API Developer Guide is live!
HELLO?? CAN YOU HEAR ME????
More times than I'm proud to admit did I utter these words into my laptop over the past few days 😅
More times than I'm proud to admit did I utter these words into my laptop over the past few days 😅
December 26, 2024 at 10:15 AM
HELLO?? CAN YOU HEAR ME????
More times than I'm proud to admit did I utter these words into my laptop over the past few days 😅
More times than I'm proud to admit did I utter these words into my laptop over the past few days 😅
𝗔 𝗱𝗲𝘃𝗲𝗹𝗼𝗽𝗲𝗿 𝗴𝘂𝗶𝗱𝗲 𝗳𝗼𝗿 𝗚𝗲𝗺𝗶𝗻𝗶’𝘀 𝗠𝘂𝗹𝘁𝗶𝗺𝗼𝗱𝗮𝗹 𝗟𝗶𝘃𝗲 𝗔𝗣𝗜
December 24, 2024 at 9:36 AM
𝗔 𝗱𝗲𝘃𝗲𝗹𝗼𝗽𝗲𝗿 𝗴𝘂𝗶𝗱𝗲 𝗳𝗼𝗿 𝗚𝗲𝗺𝗶𝗻𝗶’𝘀 𝗠𝘂𝗹𝘁𝗶𝗺𝗼𝗱𝗮𝗹 𝗟𝗶𝘃𝗲 𝗔𝗣𝗜
Developers are loving the Gemini 2.0 Multimodal Live API - we see so many of you starting to build with it 🤗
To get started with the API I wrote a small Python script (83 lines of code) that demonstrates how to set up a real-time, two-way audio communication with a Gemini language model.
To get started with the API I wrote a small Python script (83 lines of code) that demonstrates how to set up a real-time, two-way audio communication with a Gemini language model.
December 23, 2024 at 1:10 PM
Developers are loving the Gemini 2.0 Multimodal Live API - we see so many of you starting to build with it 🤗
To get started with the API I wrote a small Python script (83 lines of code) that demonstrates how to set up a real-time, two-way audio communication with a Gemini language model.
To get started with the API I wrote a small Python script (83 lines of code) that demonstrates how to set up a real-time, two-way audio communication with a Gemini language model.
December 19, 2024 at 10:04 PM
Math puzzle contest!
Gemini 2.0 Flash Thinking vs GPT-4o vs Claude 3.5 Sonnet. I was honestly surprised by the results.
Would love if someone could check with o1(Pro) 🤗
(Credit to the Bluesky community where I saw this puzzle a few days ago)
Gemini 2.0 Flash Thinking vs GPT-4o vs Claude 3.5 Sonnet. I was honestly surprised by the results.
Would love if someone could check with o1(Pro) 🤗
(Credit to the Bluesky community where I saw this puzzle a few days ago)
December 19, 2024 at 9:39 PM
Math puzzle contest!
Gemini 2.0 Flash Thinking vs GPT-4o vs Claude 3.5 Sonnet. I was honestly surprised by the results.
Would love if someone could check with o1(Pro) 🤗
(Credit to the Bluesky community where I saw this puzzle a few days ago)
Gemini 2.0 Flash Thinking vs GPT-4o vs Claude 3.5 Sonnet. I was honestly surprised by the results.
Would love if someone could check with o1(Pro) 🤗
(Credit to the Bluesky community where I saw this puzzle a few days ago)
Gemini 2.0 Flash Thinking released!
You thought we were done shipping, am I right? But the Google DeepMind folks had one more ace up their sleeves, and it's a big one!
#google #gemini #gemini2.0 #deepmind
You thought we were done shipping, am I right? But the Google DeepMind folks had one more ace up their sleeves, and it's a big one!
#google #gemini #gemini2.0 #deepmind
December 19, 2024 at 8:57 PM
Gemini 2.0 Flash Thinking released!
You thought we were done shipping, am I right? But the Google DeepMind folks had one more ace up their sleeves, and it's a big one!
#google #gemini #gemini2.0 #deepmind
You thought we were done shipping, am I right? But the Google DeepMind folks had one more ace up their sleeves, and it's a big one!
#google #gemini #gemini2.0 #deepmind
multimodal AI applications require a different architecture to make them secure
#gemini #multimodal #gemini2.0 #websockets #ai #genai #security
#gemini #multimodal #gemini2.0 #websockets #ai #genai #security
December 16, 2024 at 10:58 AM
multimodal AI applications require a different architecture to make them secure
#gemini #multimodal #gemini2.0 #websockets #ai #genai #security
#gemini #multimodal #gemini2.0 #websockets #ai #genai #security
a SIGNIFICANT number of the 30M+ users asked for this, i can assure you 😅
December 13, 2024 at 4:40 PM
a SIGNIFICANT number of the 30M+ users asked for this, i can assure you 😅
anyone else getting a warm fuzzy feeling when load_dotenv responds with `True` - or is that just me? 😅
December 13, 2024 at 10:28 AM
anyone else getting a warm fuzzy feeling when load_dotenv responds with `True` - or is that just me? 😅
yup - the ability to start prototyping with the dev API and then switch seamlessly to Vertex for production is very neat 🙂
December 12, 2024 at 9:45 AM
yup - the ability to start prototyping with the dev API and then switch seamlessly to Vertex for production is very neat 🙂
high hopes for Gemini 2.0 Pro then 😃
December 12, 2024 at 9:03 AM
high hopes for Gemini 2.0 Pro then 😃
Not sure how much to trust the judgement of an AI reporter who thought NotebookLM audio overviews were ridiculous …
December 12, 2024 at 5:28 AM
Not sure how much to trust the judgement of an AI reporter who thought NotebookLM audio overviews were ridiculous …
looks like LMArena folks liked the new Gemini 2.0 Flash model
December 11, 2024 at 7:44 PM
looks like LMArena folks liked the new Gemini 2.0 Flash model
December 11, 2024 at 4:11 PM
Gemini 2.0 Flash Experimental performs at a level comparable to that of Gemini 1.5 Pro.
December 11, 2024 at 3:51 PM
Gemini 2.0 Flash Experimental performs at a level comparable to that of Gemini 1.5 Pro.
Google introduces unified SDK for both the Gemini Developer API and the Gemini API on Vertex AI.
To be honest, while Gemini 2.0 might be the headline feature, this is actually amazing news for developers! It has been requested for quite some time and the team has finally delivered it 😍
To be honest, while Gemini 2.0 might be the headline feature, this is actually amazing news for developers! It has been requested for quite some time and the team has finally delivered it 😍
December 11, 2024 at 3:21 PM
Google introduces unified SDK for both the Gemini Developer API and the Gemini API on Vertex AI.
To be honest, while Gemini 2.0 might be the headline feature, this is actually amazing news for developers! It has been requested for quite some time and the team has finally delivered it 😍
To be honest, while Gemini 2.0 might be the headline feature, this is actually amazing news for developers! It has been requested for quite some time and the team has finally delivered it 😍
Gemini 2.0 Flash has been released - you can find it on the Gemini app and also in Google's AI Studio.
In AI Studio you can also find more details about the model:
- Pricing: 0$
- Multimodal understanding and generation
- Native tool use
- Processes 10,000 lines of code
- Knowledge cutoff Aug 2024
In AI Studio you can also find more details about the model:
- Pricing: 0$
- Multimodal understanding and generation
- Native tool use
- Processes 10,000 lines of code
- Knowledge cutoff Aug 2024
December 11, 2024 at 3:13 PM
Gemini 2.0 Flash has been released - you can find it on the Gemini app and also in Google's AI Studio.
In AI Studio you can also find more details about the model:
- Pricing: 0$
- Multimodal understanding and generation
- Native tool use
- Processes 10,000 lines of code
- Knowledge cutoff Aug 2024
In AI Studio you can also find more details about the model:
- Pricing: 0$
- Multimodal understanding and generation
- Native tool use
- Processes 10,000 lines of code
- Knowledge cutoff Aug 2024