Peter van der Putten

AI & Ax in art, business & science. Director AI Lab @Pega & researcher, @LIACS, @UniLeidenNews. @[email protected] @petervanderputten.bsky.social

[Peter van der Putten @PetervanderP]

Letters from Nature: Climate change is killing our planet, but nature's ice caps, islands and coral reefs did not have a voice. Until we asked #AI (#GPT3) to write letters to world leaders on her behalf. Poetic collab with Jeroen van der Most @Artist_Most https://t.co/sL5sd3dqh9

[Anthropic @AnthropicAI]

Our internal data shows Claude is accelerating AI development—a possible path to recursive self-improvement, or AI autonomously building a more capable successor.

It’s happening faster than we thought, and the implications deserve greater attention. https://t.co/OVVPJO7VQx

[Peter van der Putten @PetervanderP]

@AndrewYNg @SpirosMargaris Agree there is plenty room for growth across all AI roles. Also, there are plenty of platforms out there (incl us at @pega) that abstract away from the GenAI services layer making them swappable, and focus on agentic, non agentic AI and work orchestration & app layer instead

[Peter van der Putten @PetervanderP]

@bythyag I think that there is more to AI than agentic AI, we need to connect agentic AI to very predictable tools such as workflows, business rules etc not some open ended terminal process, and we need to be able to seamlessly blend agentic process, straight through proces and human work

[Matthieu wyart @MatthieuWyart]

LLMs learn by predicting tokens. World models (JEPA, data2vec) learn by predicting their own abstractions. Which needs more data? For data with hidden hierarchy, we prove the gap is exponential. https://t.co/r2uuX0lBCu

[Peter van der Putten @PetervanderP]

Here is an example of our own where we use #synthetic reference data for #RAG evaluation. https://t.co/kHON4qWtWM

[Peter van der Putten @PetervanderP]

The age of artificial artificial artificial intelligence is coming - still love this title. That said there is true benefit in using synthetic data to supplement ground truth reference data (sample link in comment). #aaai #aai #ai #synthetic https://t.co/fdcmSniQca

[a16z @a16z]

OpenAI and Anthropic are effectively telling the market they can't solve every problem with a generic AI coworker.

You don't pour billions into massive forward-deployed joint ventures if you think the next model release is going to take care of it.

In the cloud supercycle, semis led and software followed (and you didn't need Qualcomm or ARM to tell you the value was migrating up the stack).

In AI, the infra layer itself is telling us the application layer is a separate, massive opportunity they can't fully capture.

a16z's @joeschmidtiv on why the app layer isn't dead: https://t.co/84QN5Mj9T3

[Peter van der Putten @PetervanderP]

@a16z Agents need a lot of scaffolding and you also need to add case management, workflow, Ux, integration, non generative AI, business rules - the lot. And seamlessly switch between agentic, straight through process and human work. In other words a platform like @pega.

[Serena Ge (Datacurve) @serenaa_ge]

Today we’re releasing DeepSWE, a new standard for agentic coding benchmarks.

On public leaderboards, top models often look relatively close in capability. DeepSWE shows where they actually diverge, reflecting the realistic experience of developers in their day-to-day work.

[Pope Leo XIV @Pontifex]

In the era of #ArtificialIntelligence, when human dignity is threatened by new forms of dehumanization, ours is the pressing duty to remain profoundly human. We must lovingly safeguard the grandeur of humanity bestowed upon us and revealed in its fullness in Christ, the splendor of which no machine can ever replace. #MagnificaHumanitas
https://t.co/6i9MWs6LJl

[Peter van der Putten @PetervanderP]

@DanielMiessler This why agents need predictable tools like workflows and business rules. Nothing new, see our PoV and platform https://t.co/T649XRrz00

[Peter van der Putten @PetervanderP]

@rohanpaul_ai Agent harnesses may seem to appear out of nowhere but they nicely align with multi-agent systems as well as the agentic LLM / LLM-powered agents research of recent years, see our survey at https://t.co/SWnwvD6aub

[Google Gemini @GeminiApp]

Gemini Omni is here, and we’ve been seeing amazing creations all week. Here are some standouts 👇

[Peter van der Putten @PetervanderP]

Robert Downes. Cybernetics may solve tomorrow's problems mechanically. Sunday Times (Perth, WA), Sun 27 Nov 1949 #cybernetics https://t.co/t4bIB2T65P

[Peter van der Putten @PetervanderP]

1949: In such an age his thinking would be done for him by a few expert human brains and a legion of mechanical brains. These last would not be mere machines, applying power instead of human muscle, but devices able to remember, predict, analyse, deduce, take appropriate action

[Peter van der Putten @PetervanderP]

A #LLM based #chatbot experiment that tajes the same time as to cook an #egg, join in and help one of my student researchers https://t.co/1Y68azyMVW

[Peter van der Putten @PetervanderP]

@FrankRHutter @prior_labs @SAP Congrats @FrankRHutter, exciting developments

[Peter van der Putten @PetervanderP]

@akshay_pachaar The harness could and should be interpreted more broadly. What do you need when humans, regular automation, agents collaborate on reaching a business outcome? Areas such as case managements have worked that problem already for a long time.

[Peter van der Putten @PetervanderP]

@pitdesi True, thats why some gross unit margin numbers would be better. But I doubt thats favorable to OpenAI with their much larger proportion of free users.