ChatGPT - TechTalks

How ChatGPT Search affects the broader AI landscape

Ben Dickson — Thu, 31 Oct 2024 21:22:10 +0000

ChatGPT can now search the web when generating its responses. This will have implications for OpenAI and other AI companies.

The post How ChatGPT Search affects the broader AI landscape first appeared on TechTalks.

OpenAI could undercut Microsoft with new ChatGPT app for Windows

Ben Dickson — Fri, 18 Oct 2024 17:44:39 +0000

A native ChatGPT app for Windows can come at the expense of Microsoft's Copilot ecosystem.

The post OpenAI could undercut Microsoft with new ChatGPT app for Windows first appeared on TechTalks.

How to optimize ChatGPT and other LLMs for software engineering

Ben Dickson — Mon, 27 May 2024 06:33:30 +0000

A new study shows the strengths and pain points of using ChatGPT in software engineering. The can help organizations turbocharge their developers with LLMs.

The post How to optimize ChatGPT and other LLMs for software engineering first appeared on TechTalks.

How to create fine-tuned LLMs with ChatGPT

Ben Dickson — Fri, 24 May 2024 17:43:19 +0000

MonsterGPT is a ChatGPT assistant that creates fine-tuned LLMs (Mistral, Llama-2, Llama-3, Gemma, OpenELM, etc.) without requiring any coding.

The post How to create fine-tuned LLMs with ChatGPT first appeared on TechTalks.

Why you should be concerned about ChatGPT’s weird behavior

Ben Dickson — Thu, 22 Feb 2024 17:55:40 +0000

It is one thing to see ChatGPT churn out gibberish. It’s another to see a critical application fail because the underlying LLM malfunctions for unknown reasons.

The post Why you should be concerned about ChatGPT’s weird behavior first appeared on TechTalks.

Can ChatGPT write malware?

Ben Dickson — Thu, 07 Dec 2023 14:00:00 +0000

GPT-4 can help write code for ransomware. But there is much more to malware than creating the payload. And we are still safe from LLMs—for now.

The post Can ChatGPT write malware? first appeared on TechTalks.

No-code retrieval augmented generation (RAG) with LlamaIndex and ChatGPT

Ben Dickson — Wed, 22 Nov 2023 14:00:00 +0000

Retrieval augmented generation (RAG) enables you to use custom documents with LLMs to improve their precision.

The post No-code retrieval augmented generation (RAG) with LlamaIndex and ChatGPT first appeared on TechTalks.

How OpenAI uses network effects to protect ChatGPT’s market

Ben Dickson — Fri, 10 Nov 2023 14:02:27 +0000

ChatGPT is facing severe competition from other closed- and open-source LLMs. Here is how OpenAI is using network effects to solidify its market share.

The post How OpenAI uses network effects to protect ChatGPT’s market first appeared on TechTalks.

How to fine-tune GPT-3.5 or Llama 2 with a single instruction

Ben Dickson — Fri, 03 Nov 2023 14:00:00 +0000

gpt-llm-trainer is a tool that can fine-tune GPT-3.5 and Llama 2 LLMs with a single instruction.

The post How to fine-tune GPT-3.5 or Llama 2 with a single instruction first appeared on TechTalks.

Can ChatGPT self-explain its predictions?

Ben Dickson — Mon, 23 Oct 2023 13:00:00 +0000

A new study by USCC and MIT helps better understand the self-explainability capabilities of LLMs such as ChatGPT.

The post Can ChatGPT self-explain its predictions? first appeared on TechTalks.

The implications of ChatGPT’s new fine-tuning feature

Ben Dickson — Mon, 28 Aug 2023 13:00:00 +0000

OpenAI's new fine-tuning feature for GPT-3.5 Turbo is a reflection of the fast-evolving market for large language models (LLM).

The post The implications of ChatGPT’s new fine-tuning feature first appeared on TechTalks.

The truth about ChatGPT’s degrading capabilities

Ben Dickson — Mon, 24 Jul 2023 13:00:00 +0000

A new paper shows that the behavior of ChatGPT changes over time. But it doesn't necessarily mean that the model degrades.

The post The truth about ChatGPT’s degrading capabilities first appeared on TechTalks.

How to write better code with ChatGPT

Ben Dickson — Mon, 26 Jun 2023 13:00:00 +0000

ChatGPT is not a replacement for coding skills. But if used right, it can amplify your programming immensely.

The post How to write better code with ChatGPT first appeared on TechTalks.

How to reduce the costs of using ChatGPT and GPT-4

Ben Dickson — Mon, 12 Jun 2023 13:00:00 +0000

A recent study by Stanford University introduces techniques to considerably reduce the costs of using GPT-4, ChatGPT, and other LLM APIs.

The post How to reduce the costs of using ChatGPT and GPT-4 first appeared on TechTalks.

Why you should be careful about LLMs that imitate ChatGPT

Ben Dickson — Mon, 05 Jun 2023 13:00:00 +0000

Model imitation enables small LLMs to copy the style of large language models such as ChatGPT. But it also has some important limits.

The post Why you should be careful about LLMs that imitate ChatGPT first appeared on TechTalks.

How to customize LLMs like ChatGPT with your own data and documents

Ben Dickson — Mon, 01 May 2023 13:00:00 +0000

ChatGPT and other LLMs are limited to their training data. Here's how you can customize them with embeddings and your own documents.

The post How to customize LLMs like ChatGPT with your own data and documents first appeared on TechTalks.

A look at open-source alternatives to ChatGPT

Ben Dickson — Mon, 17 Apr 2023 13:00:00 +0000

All you need to know about LLaMA, Alpaca, Vicuna, and other open-source alternatives to ChatGPT, OpenAI's flagship large language model (LLM).

The post A look at open-source alternatives to ChatGPT first appeared on TechTalks.

ChatGPT is impressive, but it may slow the emergence of AGI

Charles Simon — Thu, 13 Apr 2023 13:00:00 +0000

ChatGPT represents a major advance in self-learning AI. But to make the leap to AGI, researchers must shift their focus to biologically plausible systems.

The post ChatGPT is impressive, but it may slow the emergence of AGI first appeared on TechTalks.

ChatGPT: It’s only words

Herbert Roitblat — Thu, 09 Mar 2023 14:00:00 +0000

In the rising excitement about large language models, particularly ChatGPT, people are taking sides. What is largely missing is any critical analysis of what these failures and successes tell us about the relationship between these models and intelligence.

The post ChatGPT: It’s only words first appeared on TechTalks.

Can you trust ChatGPT and other LLMs in math?

Ben Dickson — Mon, 06 Mar 2023 14:00:00 +0000

Large language models like ChatGPT are inconsistent at math. Here's why it matters.

The post Can you trust ChatGPT and other LLMs in math? first appeared on TechTalks.

ChatGPT: It can tell but does not know

Rich Heimann — Thu, 02 Mar 2023 14:00:00 +0000

ChatGPT is Polanyi's paradox in reverse. It can tell, but it doesn't know.

The post ChatGPT: It can tell but does not know first appeared on TechTalks.

What is reinforcement learning from human feedback (RLHF)?

Ben Dickson — Mon, 16 Jan 2023 14:00:00 +0000

Reinforcement learning from human feedback (RLHF) is the technique that has made ChatGPT very impressive. But there is more to RLHF that large language models (LLM).

The post What is reinforcement learning from human feedback (RLHF)? first appeared on TechTalks.

Why ChatGPT is not a threat to Google Search

Ben Dickson — Mon, 02 Jan 2023 14:00:00 +0000

ChatGPT is a remarkable LLM, with potential applications for online search. But it might be a bit of a stretch to say that it will dethrone Google.

The post Why ChatGPT is not a threat to Google Search first appeared on TechTalks.

What to (not) expect from OpenAI’s ChatGPT

Ben Dickson — Mon, 05 Dec 2022 14:00:00 +0000

OpenAI's ChatGPT, with all its successes and failures, is a reflection of the short but rich history of large language models (LLM).

The post What to (not) expect from OpenAI’s ChatGPT first appeared on TechTalks.